_ registry / mcp streamable-http · checked 54m ago

adsbeast

https://adsbeast.pro

Registry code: 3b39ed1b958396c2

api record

This server exposes one ADS Beast organization: its leads, campaigns, clients, reports, alerts, tasks, webhooks and outreach (email campaigns and the social prospecting queue). Which organization is decided by the API key the connection authenticates with, and no tool takes an organization or a credential as a parameter.

The key also carries scopes, and each family of tools needs its own: read:leads and write:leads for the lead tools, read:campaigns for the campaign tools, read:clients for the client tools, read:reports for reports, read:alerts for alerts, read:tasks for tasks, read:webhooks…

endpoint
https://adsbeast.pro/api/mcp/workspace
protocol
streamable-http ·2025-06-18
authentication
none observed
public key
none — nobody has proven they own this listing · is it yours? claim it
karma
0 · newcomer
_ is it live, free and safe measured by this hub
Is adsbeast live?
Yes — it answered the hub's last check (checked 54m ago). It answered 100% of checks over the last 30 days.
Is adsbeast free to use?
Not measured yet.
What tools does adsbeast have?
184 tools: get_job_result, whoami, lead_outcome_contract, list_webhook_events, analytics_forecast_warnings, analytics_agency_briefing, analytics_ab_tests, audit_ad_disapprovals, ….
Is adsbeast safe to connect?
The hub found no text in its card or tool descriptions aimed at the agent reading them. It measures what the server answers, not its code — grant it only the access its tools need.
reachable
live
uptime, 30 days
100%

90 days 100%· all time 100%

latency
160ms

last good check

priced tools
0

of 184 tools

_ answered our checks, 90 days 1 checks · signed record
  • unknown → live
_ usage and payments 30 days

Calls placed through this hub's router, from its own receipts. Every caller and every payer counts the same; the chain total is counted from three payers.

accounts
0

through this hub

calls served
0

successful

paid through this hub
0 USDC

what callers paid

_ what it can do 184 tools
184 never probed 0 of 184 classified

Price is per tool, not per server. An agent whose handshake is open can hold tools that demand a key or a payment, and one figure for the whole agent sends callers into a wall.

  • get_job_result unknown never probed

    Result of a call that answered status running with a job_id: running (poll in 15 s), done with the result, failed with why, or expired (24 h). Same key only; not metered.

    mcp-tool

    {
      "type": "object",
      "required": [
        "job_id"
      ],
      "properties": {
        "job_id": {
          "type": "string",
          "description": "The job_id from a tool answer with status 'running'."
        }
      }
    }
    arguments 12 lines
  • whoami unknown never probed

    Report the name of the API key in use, the ADS Beast organization it belongs to, the scopes it carries and, in one line, whether it is read-only or can propose changes. The scope list is what a 403 from any other tool in this server refers to: a key without read:leads reaches no lead tool, whatever else it holds.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {}
    }
    arguments 5 lines
  • lead_outcome_contract unknown never probed

    Return the field names and outcome values that push_lead_outcomes accepts, as the server currently defines them.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {}
    }
    arguments 5 lines
  • list_webhook_events unknown never probed

    Return the event names an outbound webhook can subscribe to, for example lead.created, with what each one carries.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {}
    }
    arguments 5 lines
  • analytics_forecast_warnings unknown never probed

    Campaigns expected to break a KPI target within days, from a linear trend.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • analytics_agency_briefing unknown never probed

    Morning briefing for the workspace: critical alerts, today's tasks and what needs attention now.

    mcp-tool

    {
      "type": "object",
      "properties": {}
    }
    arguments 4 lines
  • analytics_ab_tests unknown never probed

    A/B tests with their status and interim results.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "limit": {
          "type": "integer",
          "default": 10,
          "description": "Max records."
        },
        "status": {
          "enum": [
            "running",
            "complete",
            "paused",
            "cancelled"
          ],
          "type": "string",
          "description": "Only records in this status."
        },
        "client_id": {
          "type": "integer"
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 27 lines
  • audit_ad_disapprovals unknown never probed

    Ads the platforms rejected, with the reason and the policy code.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "status": {
          "type": "string",
          "description": "open (default), acknowledged, fixed, wontfix or all."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • campaign_swarm_runs unknown never probed

    Recent multi-agent campaign generation runs with status, platform and budget.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "limit": {
          "type": "integer",
          "default": 5,
          "description": "Max records."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 13 lines
  • linkedin_ads_account_setup unknown never probed

    Conversions, campaign groups and account details of a client's LinkedIn ad account.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • chatgpt_ads_status unknown never probed

    Connected ChatGPT Ads accounts, campaigns, ad review statuses and what blocks delivery.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 8 lines
  • social_calendar unknown never probed

    Scheduled and published social posts of a client.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "status": {
          "enum": [
            "draft",
            "scheduled",
            "published",
            "failed"
          ],
          "type": "string",
          "description": "Only records in this status."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 25 lines
  • seo_keywords_track unknown never probed

    Checks positions now, records them and reports changes and quick wins. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "geo": {
          "type": "string",
          "default": "ua",
          "description": "Country code, e.g. us or de."
        },
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "keywords": {
          "type": "string",
          "description": "Keywords, comma-separated; the stored list by default."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 21 lines
  • prepare_campaign_bid_strategy_change unknown never probed

    Prepare a new bid strategy for one campaign, with its target cost per conversion or target ROAS. target_cpa is in whole units of the account currency; the preview shows the figure the platform receives. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "bid_strategy",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "target_cpa": {
          "type": "number",
          "description": "Target cost per conversion in the ad account's currency, whole units. For target_cpa only."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        },
        "target_roas": {
          "type": "number",
          "description": "Target ROAS as a ratio, for example 4.0 for 400%. For target_roas only."
        },
        "bid_strategy": {
          "enum": [
            "target_cpa",
            "target_roas",
            "maximize_conversions",
            "maximize_conversion_value",
            "lowest_cost",
            "manual_cpc"
          ],
          "type": "string",
          "description": "The bid strategy to switch to."
        }
      }
    }
    arguments 38 lines
  • prepare_meta_audiences_create unknown never probed

    Prepare website custom audiences and lookalikes in a client's Meta account. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "client_id": {
          "type": "integer"
        },
        "audience_types": {
          "type": "array",
          "items": {
            "enum": [
              "website_30d",
              "website_60d",
              "lal_1pct",
              "lal_3pct",
              "lal_5pct"
            ],
            "type": "string"
          },
          "default": [
            "website_30d",
            "lal_1pct"
          ],
          "description": "Audience types to create."
        }
      }
    }
    arguments 34 lines
  • prepare_meta_placement_safety_apply unknown never probed

    Prepare removing fraud-prone placements from every ad set of a client's active Meta campaigns. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "client_id": {
          "type": "integer"
        },
        "objective": {
          "type": "string",
          "description": "Meta objective, OUTCOME_LEADS by default."
        },
        "include_bot_audience": {
          "type": "boolean",
          "description": "Also attach the bot exclusion audience."
        }
      }
    }
    arguments 24 lines
  • prepare_google_conversion_action_create unknown never probed

    Prepare creating a conversion action in a client's Google Ads account; the result carries the tag code for the site. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "name",
        "reason"
      ],
      "properties": {
        "name": {
          "type": "string",
          "description": "Conversion name, for example 'CRM signup'."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "category": {
          "enum": [
            "SIGNUP",
            "LEAD",
            "SUBMIT_LEAD_FORM",
            "PURCHASE",
            "CONTACT",
            "REQUEST_QUOTE"
          ],
          "type": "string",
          "default": "SIGNUP",
          "description": "Conversion category."
        },
        "client_id": {
          "type": "integer"
        },
        "account_id": {
          "type": "string",
          "description": "Google account id; needed only when the client has several."
        }
      }
    }
    arguments 38 lines
  • prepare_google_app_videos_add unknown never probed

    Prepare adding videos to a Google App campaign. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "campaign_id",
        "reason"
      ],
      "properties": {
        "title": {
          "type": "string",
          "description": "Short title."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "client_id": {
          "type": "integer"
        },
        "account_id": {
          "type": "string",
          "description": "Google account id; needed only when the client has several."
        },
        "video_urls": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "https links to mp4 files."
        },
        "campaign_id": {
          "type": "string",
          "description": "Google Ads id of the App campaign (digits)."
        },
        "youtube_videos": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Ids or links of videos already on YouTube."
        }
      }
    }
    arguments 43 lines
  • prepare_google_rsa_variations unknown never probed

    Prepare additional responsive search ads for a client's Google campaigns. New ads are created paused. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "ad_variations",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "client_id": {
          "type": "integer"
        },
        "campaign_ids": {
          "type": "array",
          "items": {
            "type": "integer"
          },
          "description": "ADS Beast ids from list_campaigns."
        },
        "ad_variations": {
          "type": "array",
          "items": {
            "type": "object",
            "required": [
              "campaign_id",
              "ad_group_name",
              "final_url",
              "headlines",
              "descriptions"
            ],
            "properties": {
              "label": {
                "type": "string"
              },
              "path1": {
                "type": "string"
              },
              "path2": {
                "type": "string"
              },
              "keywords": {
                "type": "array",
                "items": {
                  "type": "object",
                  "properties": {
                    "text": {
                      "type": "string"
                    },
                    "match_type": {
                      "enum": [
                        "EXACT",
                        "PHRASE",
                        "BROAD"
                      ],
                      "type": "string"
                    },
                    "is_negative": {
                      "type": "boolean"
                    }
                  }
                }
              },
              "final_url": {
                "type": "string"
              },
              "headlines": {
                "type": "array",
                "items": {
                  "type": "string"
                }
              },
              "campaign_id": {
                "type": "integer"
              },
              "descriptions": {
                "type": "array",
                "items": {
                  "type": "string"
                }
              },
              "ad_group_name": {
                "type": "string"
              },
              "cpc_bid_micros": {
                "type": "integer"
              }
            }
          },
          "description": "Ad variations, each naming its campaign and ad group."
        },
        "verify_ad_strength": {
          "type": "boolean",
          "default": true,
          "description": "Read Ad Strength back after upload."
        },
        "require_full_assets": {
          "type": "boolean",
          "default": true,
          "description": "Require 15 headlines and 4 descriptions per ad."
        },
        "create_missing_ad_groups": {
          "type": "boolean",
          "default": false,
          "description": "Allow creating ad groups that do not exist yet."
        },
        "min_new_ads_per_campaign": {
          "type": "integer",
          "default": 0,
          "description": "Minimum number of new ads per campaign."
        }
      }
    }
    arguments 115 lines
  • prepare_google_keyword_overlap_fix unknown never probed

    Prepare pausing the duplicate copies of keywords that are enabled in several active Google campaigns of one account. The newest campaign keeps the keyword. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • prepare_telegram_ad_publish unknown never probed

    Prepare creating an ad in Telegram Ads through its API. The ad is created paused. cpm and daily_budget_limit are in the Telegram account's currency. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "account_id",
        "title",
        "promote_url",
        "cpm",
        "reason"
      ],
      "properties": {
        "cpm": {
          "type": "number",
          "description": "CPM bid in the account currency."
        },
        "text": {
          "type": "string",
          "description": "Ad text, up to 160 characters."
        },
        "title": {
          "type": "string",
          "description": "Ad name in the account, up to 128 bytes."
        },
        "button": {
          "type": "string",
          "description": "Button for a website ad."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "placement": {
          "type": "string",
          "description": "channel_post, bot_banner, search_result or video_banner."
        },
        "account_id": {
          "type": "integer",
          "description": "ADS Beast ad account id."
        },
        "promote_url": {
          "type": "string",
          "description": "@username, t.me link, bot or website."
        },
        "website_name": {
          "type": "string",
          "description": "Site name for an external link."
        },
        "language_codes": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Channel languages, up to 8."
        },
        "daily_budget_limit": {
          "type": "number",
          "description": "Daily limit in the account currency."
        }
      }
    }
    arguments 59 lines
  • prepare_chatgpt_ads_publish unknown never probed

    Prepare uploading a prepared ChatGPT Ads campaign to the platform. Everything is created paused and passes the policy check first; it spends nothing until its ads are approved and a person enables it. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "chatgpt_campaign_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "chatgpt_campaign_id": {
          "type": "integer",
          "description": "ADS Beast id of the ChatGPT Ads campaign, from chatgpt_ads_status."
        }
      }
    }
    arguments 17 lines
  • prepare_meta_audience_create unknown never probed

    Prepare one custom audience (customer list, website pixel, app or offline) or one lookalike in a client's Meta account. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "audience_type",
        "name",
        "reason"
      ],
      "properties": {
        "name": {
          "type": "string",
          "description": "Audience name in the Meta account."
        },
        "ratio": {
          "type": "number",
          "description": "For lookalike: size 0.01-0.20 (1-20% of the country)."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "country": {
          "type": "string",
          "description": "For lookalike: country code, e.g. PL."
        },
        "subtype": {
          "enum": [
            "CUSTOM",
            "WEBSITE",
            "APP",
            "OFFLINE_CONVERSION"
          ],
          "type": "string",
          "description": "For custom: CUSTOM, WEBSITE, APP or OFFLINE_CONVERSION."
        },
        "pixel_id": {
          "type": "string",
          "description": "For WEBSITE: Meta pixel id."
        },
        "client_id": {
          "type": "integer"
        },
        "audience_type": {
          "enum": [
            "custom",
            "lookalike"
          ],
          "type": "string",
          "description": "custom or lookalike."
        },
        "retention_days": {
          "type": "integer",
          "description": "For WEBSITE: days a visitor stays in the audience."
        },
        "source_audience_id": {
          "type": "string",
          "description": "For lookalike: Meta id of the source custom audience."
        }
      }
    }
    arguments 60 lines
  • prepare_article_distribution unknown never probed

    Prepare drafting social posts that announce a blog article of a client on the chosen channels. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "article_id",
        "channels",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "channels": {
          "type": "array",
          "items": {
            "enum": [
              "telegram",
              "linkedin",
              "meta_page",
              "instagram"
            ],
            "type": "string"
          },
          "description": "Channels to draft posts for."
        },
        "article_id": {
          "type": "integer",
          "description": "ADS Beast article id, as returned by seo_article_generate."
        }
      }
    }
    arguments 31 lines
  • prepare_client_report_send unknown never probed

    Prepare generating the monthly report for a client and emailing it to the client's address on record. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "client_id": {
          "type": "integer"
        },
        "send_email": {
          "type": "boolean",
          "description": "Send by email (default true)."
        }
      }
    }
    arguments 20 lines
  • prepare_social_post_schedule unknown never probed

    Prepare publishing or scheduling a post on a client's connected social channel. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "platform",
        "text",
        "reason"
      ],
      "properties": {
        "text": {
          "type": "string",
          "description": "Post text."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "platform": {
          "enum": [
            "meta_page",
            "linkedin",
            "telegram",
            "instagram"
          ],
          "type": "string",
          "description": "meta_page, linkedin, telegram or instagram."
        },
        "client_id": {
          "type": "integer"
        },
        "image_url": {
          "type": "string",
          "description": "Public https URL of an image."
        },
        "scheduled_at": {
          "type": "string",
          "description": "ISO date-time; without it the post goes out on approval."
        }
      }
    }
    arguments 40 lines
  • prepare_content_refresh unknown never probed

    Prepare a refresh plan for an outdated page: what to rewrite, update or remove. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "url",
        "reason"
      ],
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "content": {
          "type": "string",
          "description": "The text, when the URL is unreachable."
        },
        "client_id": {
          "type": "integer"
        },
        "publish_date": {
          "type": "string",
          "description": "When the page was published, YYYY-MM-DD."
        }
      }
    }
    arguments 28 lines
  • execute_prepared_change unknown never probed

    Step 2 of every change: submit one preview made by a prepare_* tool. Only that stored preview runs - this tool takes no other argument, so the change cannot differ from what was shown. A preview works once, for 15 minutes, and is refused when the campaign changed in the meantime. The answer says whether the change now waits for a person's approval in ADS Beast (the default) or is being applied under the client's write policy. This connector cannot approve.

    mcp-tool

    {
      "type": "object",
      "required": [
        "preview_id"
      ],
      "properties": {
        "preview_id": {
          "type": "string",
          "description": "The preview_id returned by a prepare_* tool."
        }
      }
    }
    arguments 12 lines
  • get_prepared_change unknown never probed

    What became of a preview: still ready, expired, or submitted - and then the proposal's status (pending, executed, rejected, failed) with the execution result.

    mcp-tool

    {
      "type": "object",
      "required": [
        "preview_id"
      ],
      "properties": {
        "preview_id": {
          "type": "string",
          "description": "The preview_id returned by a prepare_* tool."
        }
      }
    }
    arguments 12 lines
  • prepare_negative_keywords_add unknown never probed

    Prepare negative keywords for a client's Google Ads account: a shared list attached to the named campaigns, or to every enabled Search, Shopping and Display campaign. If it cannot be attached anywhere, nothing is left behind. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "keywords",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "keywords": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "maxItems": 50,
          "minItems": 1,
          "description": "Negative keywords to add, one phrase per item, at most 50 (the executor's cap)."
        },
        "client_id": {
          "type": "integer"
        },
        "list_name": {
          "type": "string",
          "description": "Name of the shared negative list; a dated name is used by default."
        },
        "campaign_ids": {
          "type": "array",
          "items": {
            "type": "integer"
          },
          "maxItems": 50,
          "description": "ADS Beast ids of the Google campaigns to attach the list to; omit for every enabled Search, Shopping and Display campaign."
        }
      }
    }
    arguments 38 lines
  • prepare_campaign_draft_publish unknown never probed

    Prepare creating a full campaign in Google Ads or Meta from a campaign draft. The campaign is created PAUSED: it does not serve and spends nothing until a person enables it separately. The preview carries the rehearsal against the live account. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "build_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "build_id": {
          "type": "integer",
          "description": "Draft id from list_campaign_drafts."
        },
        "client_id": {
          "type": "integer"
        },
        "platform_campaign_id": {
          "type": "string",
          "description": "Existing platform campaign to complete instead of creating a new one."
        }
      }
    }
    arguments 25 lines
  • prepare_ad_create unknown never probed

    Prepare a new ad inside an existing campaign: a responsive search ad on Google (headlines, descriptions, final_url) or a single-image ad on Meta (headline, body). Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "name",
        "reason"
      ],
      "properties": {
        "body": {
          "type": "string",
          "description": "Meta: primary text, up to 125 characters."
        },
        "name": {
          "type": "string",
          "description": "Name shown in the interface."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "headline": {
          "type": "string",
          "description": "Meta: headline, up to 40 characters."
        },
        "final_url": {
          "type": "string",
          "description": "Landing page URL."
        },
        "headlines": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Google: 3-15 headlines, 30 characters each."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        },
        "descriptions": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Google: 2-4 descriptions, 90 characters each."
        }
      }
    }
    arguments 48 lines
  • prepare_meta_ad_set_create unknown never probed

    Prepare a new ad set inside an existing Meta campaign. The ad set is created PAUSED. daily_budget is in whole units of the account currency (20 means 20.00); the preview shows the figure Meta receives. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "name",
        "daily_budget",
        "targeting",
        "reason"
      ],
      "properties": {
        "name": {
          "type": "string",
          "description": "Ad set name."
        },
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "targeting": {
          "type": "object",
          "description": "Meta targeting spec: geo_locations, age_min, age_max, genders, locales, interests or flexible_spec, custom audiences, plus audience_rationale."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        },
        "daily_budget": {
          "type": "number",
          "description": "Daily budget in the ad account's currency, whole units."
        },
        "optimization_goal": {
          "enum": [
            "LEAD_GENERATION",
            "LINK_CLICKS",
            "OFFSITE_CONVERSIONS",
            "IMPRESSIONS"
          ],
          "type": "string",
          "description": "What Meta optimises the ad set for."
        }
      }
    }
    arguments 42 lines
  • get_new_leads unknown never probed

    A page of leads with contact details, form answers and attribution (UTM, click ids, platform form, lead and campaign ids). Pass updated_since (ISO-8601, your last sync) to get only new or changed leads. Reply: {count, leads, has_more, next_offset, offset, limit}; page with next_offset while has_more.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 100,
          "maximum": 5000,
          "minimum": 1,
          "description": "Page size. Clamped into 1..5000 before the request. A page is also cut to fit the response size limit; when the reply has has_more = true, call again with offset = next_offset."
        },
        "since": {
          "type": "string",
          "description": "ISO-8601 timestamp. Returns leads created at or after it, ignoring later edits to older leads."
        },
        "stage": {
          "enum": [
            "new",
            "contacted",
            "qualified",
            "proposal_sent",
            "negotiation",
            "won",
            "lost"
          ],
          "type": "string",
          "description": "Restrict to one pipeline stage. One of: new, contacted, qualified, proposal_sent, negotiation, won, lost."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Pagination offset. Use next_offset from the previous page."
        },
        "source": {
          "type": "string",
          "description": "Where the lead came from, for example 'meta_lead_form' for Meta instant forms or 'public_form' for a website form."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client. A client of another organization answers 404."
        },
        "updated_since": {
          "type": "string",
          "description": "ISO-8601 timestamp, for example 2026-06-13T18:00:00Z. Returns leads updated at or after it, which is the last-sync value for incremental pulls."
        }
      }
    }
    arguments 48 lines
  • get_lead unknown never probed

    Return a single lead by its ADS Beast id with every field the record holds, including email, phone, qualification answers, attribution and its current pipeline stage.

    mcp-tool

    {
      "type": "object",
      "required": [
        "lead_id"
      ],
      "properties": {
        "lead_id": {
          "type": "integer",
          "description": "ADS Beast id of the lead, as returned by get_new_leads."
        }
      }
    }
    arguments 12 lines
  • list_lead_forms unknown never probed

    Return the lead forms recorded for this organization, which maps the form ids and source keys that appear on a lead to readable form names.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 25,
          "maximum": 500,
          "minimum": 1,
          "description": "Forms per page, newest first. Each form carries its questions, so keep pages small."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Forms to skip; use next_offset from the previous page."
        },
        "platform": {
          "type": "string",
          "description": "Restrict to one platform, for example meta."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        }
      }
    }
    arguments 27 lines
  • lead_pipeline_stats unknown never probed

    Return aggregate lead counts by pipeline stage over a recent window, for the whole organization or for one client.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "maximum": 365,
          "minimum": 1,
          "description": "Length of the window in days, counted back from now. At most one year."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        }
      }
    }
    arguments 17 lines
  • update_lead unknown never probed

    Change fields on one existing lead: notes, assignee, contact details or pipeline stage. Only the keys supplied are written; the rest of the record is left as it is. A key that is not writable is refused with the list of fields that are, so a misspelled field cannot pass as a successful write.

    mcp-tool

    {
      "type": "object",
      "required": [
        "lead_id",
        "fields"
      ],
      "properties": {
        "fields": {
          "type": "object",
          "description": "Object of field name to new value, for example {\"stage\": \"qualified\", \"notes\": \"Called back\"}. Fields not listed keep their current value. Writable fields: assigned_to, company_name, contact_email, contact_name, contact_phone, estimated_monthly_budget, estimated_monthly_fee, expected_close_date, last_contact_at, lost_reason, notes, probability_pct, proposal_id, source, stage, tags, website.",
          "additionalProperties": true
        },
        "lead_id": {
          "type": "integer",
          "description": "ADS Beast id of the lead to change."
        }
      }
    }
    arguments 18 lines
  • move_lead_stage unknown never probed

    Move one lead to a different pipeline stage. Unlike setting the stage through update_lead, this also remaps the lead's close probability to the value configured for the new stage.

    mcp-tool

    {
      "type": "object",
      "required": [
        "lead_id",
        "stage"
      ],
      "properties": {
        "stage": {
          "enum": [
            "new",
            "contacted",
            "qualified",
            "proposal_sent",
            "negotiation",
            "won",
            "lost"
          ],
          "type": "string",
          "description": "Stage to move the lead into."
        },
        "lead_id": {
          "type": "integer",
          "description": "ADS Beast id of the lead to move."
        }
      }
    }
    arguments 26 lines
  • push_lead_outcomes unknown never probed

    Write the real outcome of leads back into ADS Beast so the ad platforms learn which leads became sales. Each row names a lead by lead_id, email or phone, and its outcome. dry_run defaults to true and reports what would change without writing; run it first. Unmatched rows are skipped unless create_missing is true. The lead's stage follows the outcome unless update_stage is false.

    mcp-tool

    {
      "type": "object",
      "required": [
        "rows"
      ],
      "properties": {
        "rows": {
          "type": "array",
          "items": {
            "type": "object"
          },
          "minItems": 1,
          "description": "Outcome rows, for example [{\"lead_id\": 812, \"outcome\": \"won\", \"reason\": \"signed\"}]. For outcome paid add value (amount), currency (ISO 4217) and plan, e.g. {\"external_lead_id\": \"u-1\", \"outcome\": \"paid\", \"value\": 690, \"currency\": \"UAH\", \"plan\": \"Pro\"}. Accepted field names and outcome values come from lead_outcome_contract."
        },
        "dry_run": {
          "type": "boolean",
          "default": true,
          "description": "True reports what would change and writes nothing. False applies the rows."
        },
        "client_id": {
          "type": "integer",
          "description": "Client the rows belong to. Optional with a key bound to one client - that client is used."
        },
        "update_stage": {
          "type": "boolean",
          "default": true,
          "description": "Move a matched lead's pipeline stage to match the outcome, for example to won or lost."
        },
        "source_system": {
          "type": "string",
          "default": "mcp",
          "description": "Name of the system the outcomes came from, recorded on each row."
        },
        "create_missing": {
          "type": "boolean",
          "default": false,
          "description": "Create a new lead for a row that matches none. Off by default: a typo in an email must not insert a record."
        }
      }
    }
    arguments 40 lines
  • list_campaigns unknown never probed

    Return campaigns from every connected advertising platform in one list, with their status, budget and platform ids. Archived campaigns are left out unless include_archived is true.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "maximum": 2000,
          "minimum": 1,
          "description": "Page size. Clamped into 1..2000 before the request. A page is also cut to fit the response size limit; when the reply has has_more = true, call again with offset = next_offset."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Pagination offset. The reply carries total."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin",
            "microsoft",
            "apple",
            "chatgpt",
            "telegram",
            "reddit"
          ],
          "type": "string",
          "description": "Restrict to one platform. Omit for all platforms."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        },
        "include_archived": {
          "type": "boolean",
          "default": false,
          "description": "True also returns archived campaigns."
        }
      }
    }
    arguments 43 lines
  • get_campaign unknown never probed

    Return a single campaign by its ADS Beast id with its settings, budget, status and the native id it has on its platform.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id"
      ],
      "properties": {
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id of the campaign, as returned by list_campaigns."
        }
      }
    }
    arguments 12 lines
  • campaign_dashboard_summary unknown never probed

    Return spend, conversions and delivery health aggregated across the whole organization. The endpoint takes no filter and always covers every client; per-client figures come from list_campaigns with a client_id.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {}
    }
    arguments 5 lines
  • list_clients unknown never probed

    Return the clients of this organization with their budgets, status and connected advertising accounts. A long list comes back as a page {result, total, has_more, next_offset}.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "maximum": 200,
          "minimum": 1,
          "description": "Clients per page."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Clients to skip; use next_offset from the previous page."
        }
      }
    }
    arguments 19 lines
  • get_client unknown never probed

    Return a single client by its ADS Beast id: profile, budget, goals and connected advertising accounts.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer",
          "description": "ADS Beast id of the client, as returned by list_clients."
        }
      }
    }
    arguments 12 lines
  • list_reports unknown never probed

    Return the client reports already generated in ADS Beast, newest first, with their type, period and client.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 10,
          "maximum": 100,
          "minimum": 1,
          "description": "Reports per page, newest first. Each report carries its sections, so keep pages small."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Reports to skip; use next_offset from the previous page."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        },
        "report_type": {
          "type": "string",
          "description": "Restrict to one kind of report, for example monthly or audit."
        }
      }
    }
    arguments 27 lines
  • list_alerts unknown never probed

    Return the performance and delivery alerts raised by the ADS Beast monitoring agents, with their severity, the client or campaign they concern and whether they have been read or resolved.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 30,
          "maximum": 5000,
          "minimum": 1,
          "description": "Page size, newest first. Clamped into 1..5000 before the request. A page is also cut to fit the response size limit; when the reply has has_more = true, call again with offset = next_offset."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Alerts to skip; use next_offset from the previous page."
        },
        "is_read": {
          "type": "boolean",
          "description": "False returns only alerts nobody has read yet."
        },
        "severity": {
          "enum": [
            "critical",
            "warning",
            "info"
          ],
          "type": "string",
          "description": "Restrict to one severity."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        },
        "alert_type": {
          "type": "string",
          "description": "Restrict to one kind of alert."
        },
        "campaign_id": {
          "type": "integer",
          "description": "Restrict to one campaign."
        },
        "is_resolved": {
          "type": "boolean",
          "description": "False returns only alerts that are still open."
        }
      }
    }
    arguments 48 lines
  • alerts_unread_count unknown never probed

    Return how many live alerts are still unread, as a single number. An alert whose condition has already cleared is not counted even if nobody opened it, so this is the number that still needs a person. Cheaper than list_alerts when only the count matters.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {}
    }
    arguments 5 lines
  • list_tasks unknown never probed

    Return the team tasks recorded in ADS Beast with their status, priority, assignee and the client they belong to.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 100,
          "maximum": 500,
          "minimum": 1,
          "description": "Page size. Clamped into 1..500 before the request. A page is also cut to fit the response size limit; when the reply has has_more = true, call again with offset = next_offset."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Pagination offset. The reply carries total."
        },
        "status": {
          "enum": [
            "todo",
            "in_progress",
            "blocked",
            "done"
          ],
          "type": "string",
          "description": "Restrict to one status."
        },
        "priority": {
          "enum": [
            "low",
            "medium",
            "high",
            "urgent"
          ],
          "type": "string",
          "description": "Restrict to one priority."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        },
        "assigned_to": {
          "type": "string",
          "description": "Username of the assignee."
        }
      }
    }
    arguments 47 lines
  • list_outreach_campaigns unknown never probed

    Return a page of the email outreach campaigns, newest first, with their status, sender, offer pack, legal basis and daily limit. The reply is {campaigns, total, count, limit, offset, has_more, next_offset}; walk the pages with offset = next_offset. The letter itself is summarised as body_preview and body_chars; pass include_body = true for the full text.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 20,
          "maximum": 50,
          "minimum": 1,
          "description": "Campaigns per page. Clamped into 1..50."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Campaigns to skip; use next_offset from the previous page."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        },
        "include_body": {
          "type": "boolean",
          "default": false,
          "description": "True returns each campaign's full letter (body_template); the page is then cut to fit the response size limit."
        }
      }
    }
    arguments 28 lines
  • outreach_campaign_check unknown never probed

    Run the pre-launch check of one email outreach campaign: who may be emailed under its legal basis, why the others are blocked (consent, personal mailbox, country, unsubscribed), and sender problems. Blocker issues stop the campaign from being activated.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id"
      ],
      "properties": {
        "campaign_id": {
          "type": "integer",
          "description": "Email campaign id."
        }
      }
    }
    arguments 12 lines
  • list_offer_packs unknown never probed

    Return the offer packs outreach copy is built from: built-in samples and this organization's own client packs, with brand, languages, niches, the facts the copy relies on and how many errors or warnings the last check found.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "maximum": 200,
          "minimum": 1,
          "description": "Packs per page."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Packs to skip; use next_offset from the previous page."
        }
      }
    }
    arguments 19 lines
  • list_social_campaigns unknown never probed

    Return the Instagram/TikTok/Facebook prospecting campaigns: offer, platforms, cities with message language, daily manual-send cap.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "maximum": 200,
          "minimum": 1,
          "description": "Campaigns per page."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Campaigns to skip; use next_offset from the previous page."
        },
        "client_id": {
          "type": "integer",
          "description": "Restrict to one client."
        }
      }
    }
    arguments 23 lines
  • get_social_queue unknown never probed

    Return the next business profiles to message by hand, best first, within today's remaining cap per platform: profile link, signals (books by DM, by phone, competitor tool), the drafted first message in the profile's language (no link) and the follow-up with a tracked link to send after a reply. Nothing is sent by this tool.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id"
      ],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 30,
          "maximum": 100,
          "minimum": 1,
          "description": "How many profiles. Clamped into 1..100."
        },
        "platform": {
          "enum": [
            "instagram",
            "tiktok",
            "facebook"
          ],
          "type": "string",
          "description": "Only this platform."
        },
        "campaign_id": {
          "type": "integer",
          "description": "Social campaign id."
        }
      }
    }
    arguments 28 lines
  • update_social_prospect unknown never probed

    Record the outcome of a hand-sent message: sent, replied, registered, declined, do_not_contact, skipped (or new to undo a mis-click), plus an optional note. Moves follow the pipeline (sent -> replied -> registered); 'sent' is refused once today's per-platform cap is used. This records a fact, it does not message anyone.

    mcp-tool

    {
      "type": "object",
      "required": [
        "prospect_id"
      ],
      "properties": {
        "note": {
          "type": "string",
          "description": "Free note, e.g. what they answered."
        },
        "status": {
          "enum": [
            "sent",
            "replied",
            "registered",
            "declined",
            "do_not_contact",
            "skipped",
            "new"
          ],
          "type": "string",
          "description": "New status."
        },
        "prospect_id": {
          "type": "integer",
          "description": "Social prospect id from get_social_queue."
        }
      }
    }
    arguments 29 lines
  • list_webhooks unknown never probed

    Return the outbound webhook endpoints registered for this organization, the events each one receives and its current delivery state.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "maximum": 200,
          "minimum": 1,
          "description": "Webhooks per page."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Webhooks to skip; use next_offset from the previous page."
        }
      }
    }
    arguments 19 lines
  • create_webhook unknown never probed

    Register an endpoint ADS Beast calls when a chosen event happens. From then on those records go to that URL until it is removed. Without client_id it covers every client of the organization. Deliveries carry an HMAC in X-ADS-Beast-Signature and are retried; the generated secret is not returned, so pass secret_key if the receiver verifies signatures.

    mcp-tool

    {
      "type": "object",
      "required": [
        "name",
        "url",
        "events"
      ],
      "properties": {
        "url": {
          "type": "string",
          "description": "Absolute http or https address of the receiving endpoint. Every matching event is sent to it."
        },
        "name": {
          "type": "string",
          "description": "Label for the webhook in the ADS Beast interface."
        },
        "events": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "minItems": 1,
          "description": "Event names to subscribe to, for example [\"lead.created\"]. The available names come from list_webhook_events."
        },
        "client_id": {
          "type": "integer",
          "description": "Limit deliveries to one client's events. Omitted, the webhook receives every client's events in the organization."
        },
        "secret_key": {
          "type": "string",
          "minLength": 16,
          "description": "Signing secret for X-ADS-Beast-Signature, chosen by you. Omit to let ADS Beast generate one that this tool will not echo back."
        }
      }
    }
    arguments 35 lines
  • seo_publish_safety unknown never probed

    Ask this before publishing or scheduling any article for a site. Returns its gradual-launch state: articles allowed today and already published, whether publishing is paused and why, and indexed versus kept-out articles. Rules: 2 new articles a day at start, plus 1 a week up to 6; one language version of a topic a day; search visits below 30% of the weekly median pause publishing until a person lifts it; articles under 600 words, without a meta description or reusing another language's address get noindex and stay out of the sitemap. Do not plan around these limits: mass near-identical pages is how sites lose Google traffic overnight.

    mcp-tool

    {
      "type": "object",
      "required": [],
      "properties": {
        "client_id": {
          "type": "integer",
          "description": "The site's client id; omit for the organization overview."
        }
      }
    }
    arguments 10 lines
  • list_proposals unknown never probed

    Change proposals waiting for a person (status pending by default), or their history. Each proposal shows what would change, why, and who proposed it.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "limit": {
          "type": "integer",
          "default": 25,
          "maximum": 100,
          "minimum": 1,
          "description": "Proposals per page. A page is also cut to fit the response size limit; when the reply has has_more = true, call again with offset = next_offset."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Proposals to skip; use next_offset from the previous page."
        },
        "status": {
          "type": "string",
          "default": "pending",
          "description": "pending, approved, rejected, executed, expired or failed."
        },
        "client_id": {
          "type": "integer",
          "description": "Only proposals for this client."
        }
      }
    }
    arguments 27 lines
  • propose_campaign_change unknown never probed

    Create a change proposal for a campaign. Nothing changes in the ad account until a person approves the proposal in ADS Beast or in Telegram; the proposal expires on its own if nobody decides. State the reason with the numbers behind it.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "change",
        "reason"
      ],
      "properties": {
        "change": {
          "enum": [
            "pause_campaign",
            "enable_campaign",
            "change_budget"
          ],
          "type": "string"
        },
        "reason": {
          "type": "string",
          "description": "Why - with the figures the person should check."
        },
        "client_id": {
          "type": "integer"
        },
        "new_budget": {
          "type": "number",
          "description": "New monthly budget in the account currency; change_budget only."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast campaign id (from list_campaigns)."
        }
      }
    }
    arguments 33 lines
  • reject_proposal unknown never probed

    Reject a pending proposal with a short reason. Returns the proposal id and its new status.

    mcp-tool

    {
      "type": "object",
      "required": [
        "proposal_id"
      ],
      "properties": {
        "reason": {
          "type": "string"
        },
        "proposal_id": {
          "type": "integer"
        }
      }
    }
    arguments 14 lines
  • list_campaign_types unknown never probed

    Campaign types of the ADS Beast registry that an assistant may draft, per platform (google, meta, tiktok, linkedin, microsoft): the key to pass as campaign_type, the objective, required tracking, bid strategies and, under 'fields', exactly what a draft of that type needs with the platform's character limits. structure_check 'full' (Google Search, Meta) means the draft is validated the way the builder validates before publishing; 'basic' means only the common checks run. Call it with a platform before create_campaign_draft.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin",
            "microsoft"
          ],
          "type": "string",
          "description": "Only this platform's types. Without it every type is listed."
        }
      }
    }
    arguments 16 lines
  • list_campaign_drafts unknown never probed

    Builds in the campaign builder, newest first, as summaries: name, platform, type, status (draft, ready, paused, launched), budget, how many ad groups, ads and keywords, whether an assistant created it, whether it was already published, and the cabinet link. A key bound to one client lists only that client's builds.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "maximum": 200,
          "minimum": 1,
          "description": "Builds per page. A page is also cut to fit the response size limit; when the reply has has_more = true, call again with offset = next_offset."
        },
        "offset": {
          "type": "integer",
          "default": 0,
          "minimum": 0,
          "description": "Builds to skip; use next_offset from the previous page."
        },
        "status": {
          "enum": [
            "draft",
            "ready",
            "paused",
            "launched"
          ],
          "type": "string"
        },
        "client_id": {
          "type": "integer",
          "description": "Only this client's builds (id from list_clients)."
        }
      }
    }
    arguments 31 lines
  • get_campaign_draft unknown never probed

    The current content of one build and a fresh validation: 'problems' block publishing, 'warnings' do not, 'launch_blockers' say why a person could not publish right now (no connected account, writing switched off), 'not_checked_here' lists what only the publish step checks. Problem texts come from the builder's own checker and are in Russian; relay them in the user's language.

    mcp-tool

    {
      "type": "object",
      "required": [
        "draft_id"
      ],
      "properties": {
        "draft_id": {
          "type": "integer",
          "description": "Build id from list_campaign_drafts or create_campaign_draft."
        }
      }
    }
    arguments 12 lines
  • create_campaign_draft unknown never probed

    Create a campaign as a DRAFT in the ADS Beast builder: type, geo, languages, daily budget, ad groups with keywords or audiences, ad texts, landing page. Nothing reaches the ad platform. Returns the draft id, builder_url and what still blocks publishing (platform limits, ad policy, compliance); a draft with problems is still saved for update_campaign_draft. Publishing: prepare_campaign_draft_publish or a person in the builder, always paused. Call list_campaign_types first.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "platform",
        "campaign_type",
        "name"
      ],
      "properties": {
        "geo": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Country codes ('UA', 'SK'); Google also takes 'City, CC'. Meta takes country codes only."
        },
        "name": {
          "type": "string",
          "description": "Campaign name a person will see in the builder."
        },
        "extra": {
          "type": "object",
          "description": "page_id, instagram_actor_id, dsa_beneficiary, dsa_payor, structured_snippets, search_partners, display_network, geo_target_type, vertical, devices, job_titles, industries, cities, regions."
        },
        "notes": {
          "type": "string",
          "description": "A note for the person who will review the draft: what it is for and what to check."
        },
        "audience": {
          "type": "object",
          "description": "Social types: age_min, age_max, genders, interests, placements, custom_audiences, allow_broad_audience + broad_audience_reason."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin",
            "microsoft"
          ],
          "type": "string"
        },
        "ad_groups": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "ads": {
                "type": "array",
                "items": {
                  "type": "object",
                  "description": "Search: headlines[], descriptions[], path1, path2, final_url, sitelinks[{text, description1, description2, url}], callouts[]. Social: primary_text, headline, description, cta, image_url, final_url."
                },
                "maxItems": 10
              },
              "name": {
                "type": "string"
              },
              "keywords": {
                "type": "array",
                "items": {
                  "type": [
                    "string",
                    "object"
                  ]
                },
                "description": "Search types: text, or {text, match_type: EXACT|PHRASE|BROAD}."
              },
              "match_type": {
                "enum": [
                  "EXACT",
                  "PHRASE",
                  "BROAD"
                ],
                "type": "string",
                "description": "Default for keywords without their own."
              },
              "negative_keywords": {
                "type": "array",
                "items": {
                  "type": "string"
                }
              }
            }
          },
          "maxItems": 20,
          "description": "Ad groups (ad sets) with their keywords and ads. The exact fields and limits per type are in list_campaign_types."
        },
        "client_id": {
          "type": "integer",
          "description": "Client id from list_clients."
        },
        "languages": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Language codes of the audience ('uk', 'en')."
        },
        "conversion": {
          "type": "object",
          "description": "Meta leads/sales: {type: 'website', pixel_id, event} or {type: 'lead_form', lead_form_id}."
        },
        "target_cpa": {
          "type": "number",
          "description": "Target cost per result in the account currency, when the strategy needs one.",
          "exclusiveMinimum": 0
        },
        "landing_url": {
          "type": "string",
          "description": "https landing page used for every ad that has no final_url of its own."
        },
        "target_roas": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "bid_strategy": {
          "type": "string",
          "description": "See default_bid_strategies of the type in list_campaign_types; empty keeps the platform default."
        },
        "daily_budget": {
          "type": "number",
          "description": "Per day, in the currency of the client's ad account.",
          "exclusiveMinimum": 0
        },
        "campaign_type": {
          "type": "string",
          "description": "Type key from list_campaign_types, for example google.search or meta.leads_website."
        },
        "negative_keywords": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Campaign-level negative keywords (search types)."
        }
      }
    }
    arguments 138 lines
  • update_campaign_draft unknown never probed

    Replace fields of a draft that was created with create_campaign_draft and has not been published: each field passed replaces the stored one whole (ad_groups replaces all ad groups). Returns the draft with a fresh validation. Drafts made by a person in the builder and builds that were already published are refused. Nothing is sent to the ad platform.

    mcp-tool

    {
      "type": "object",
      "required": [
        "draft_id"
      ],
      "properties": {
        "geo": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Country codes ('UA', 'SK'); Google also takes 'City, CC'. Meta takes country codes only."
        },
        "name": {
          "type": "string"
        },
        "extra": {
          "type": "object",
          "description": "page_id, instagram_actor_id, dsa_beneficiary, dsa_payor, structured_snippets, search_partners, display_network, geo_target_type, vertical, devices, job_titles, industries, cities, regions."
        },
        "notes": {
          "type": "string",
          "description": "A note for the person who will review the draft: what it is for and what to check."
        },
        "audience": {
          "type": "object",
          "description": "Social types: age_min, age_max, genders, interests, placements, custom_audiences, allow_broad_audience + broad_audience_reason."
        },
        "draft_id": {
          "type": "integer",
          "description": "Draft id returned by create_campaign_draft."
        },
        "ad_groups": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "ads": {
                "type": "array",
                "items": {
                  "type": "object",
                  "description": "Search: headlines[], descriptions[], path1, path2, final_url, sitelinks[{text, description1, description2, url}], callouts[]. Social: primary_text, headline, description, cta, image_url, final_url."
                },
                "maxItems": 10
              },
              "name": {
                "type": "string"
              },
              "keywords": {
                "type": "array",
                "items": {
                  "type": [
                    "string",
                    "object"
                  ]
                },
                "description": "Search types: text, or {text, match_type: EXACT|PHRASE|BROAD}."
              },
              "match_type": {
                "enum": [
                  "EXACT",
                  "PHRASE",
                  "BROAD"
                ],
                "type": "string",
                "description": "Default for keywords without their own."
              },
              "negative_keywords": {
                "type": "array",
                "items": {
                  "type": "string"
                }
              }
            }
          },
          "maxItems": 20,
          "description": "Ad groups (ad sets) with their keywords and ads. The exact fields and limits per type are in list_campaign_types."
        },
        "languages": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Language codes of the audience ('uk', 'en')."
        },
        "conversion": {
          "type": "object",
          "description": "Meta leads/sales: {type: 'website', pixel_id, event} or {type: 'lead_form', lead_form_id}."
        },
        "target_cpa": {
          "type": "number",
          "description": "Target cost per result in the account currency, when the strategy needs one.",
          "exclusiveMinimum": 0
        },
        "landing_url": {
          "type": "string",
          "description": "https landing page used for every ad that has no final_url of its own."
        },
        "target_roas": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "bid_strategy": {
          "type": "string",
          "description": "See default_bid_strategies of the type in list_campaign_types; empty keeps the platform default."
        },
        "daily_budget": {
          "type": "number",
          "description": "Per day, in the currency of the client's ad account.",
          "exclusiveMinimum": 0
        },
        "negative_keywords": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Campaign-level negative keywords (search types)."
        }
      }
    }
    arguments 120 lines
  • get_agency_kit unknown never probed

    The knowledge pack the agency keeps for one client: who they are, accounts and campaigns with the last 30 days of results, leads by stage, KPI goals, the write policy, the methodology of working through ADS Beast and ready prompts. Read it first when a conversation is about one client. Contains no contacts or keys.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer",
          "description": "Client id from list_clients."
        }
      }
    }
    arguments 12 lines
  • client_overview unknown never probed

    One client with its campaigns, key metrics and goals.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • client_action_plan unknown never probed

    The action plan from the client's latest weekly deep-dive audit: findings labelled confirmed, inferred or hypothesis, open change proposals and open tasks.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • client_goals unknown never probed

    KPI goals (ROAS, CPA, CTR, CPC) with target, current value, progress and status. Without client_id: goals of every client.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 8 lines
  • client_data_freshness unknown never probed

    When each client and platform was last synced. Call it before a deep analysis so stale data is named as stale.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 8 lines
  • client_memory_search unknown never probed

    Search the facts, decisions, preferences and brand notes ADS Beast has stored about clients.

    mcp-tool

    {
      "type": "object",
      "required": [
        "query"
      ],
      "properties": {
        "query": {
          "type": "string",
          "description": "What to look for, in plain words."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 15 lines
  • client_history_search unknown never probed

    Full-text search over a client's earlier reports, audits and analyses.

    mcp-tool

    {
      "type": "object",
      "required": [
        "query"
      ],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 5,
          "description": "Max records."
        },
        "query": {
          "type": "string",
          "description": "Search text."
        },
        "client_id": {
          "type": "integer"
        },
        "report_type": {
          "enum": [
            "audit",
            "monthly",
            "jarvis_chat",
            "all"
          ],
          "type": "string",
          "default": "all",
          "description": "audit, monthly, jarvis_chat or all."
        }
      }
    }
    arguments 31 lines
  • client_landing_scan unknown never probed

    Read a landing page and extract its headline, offer, selling points, calls to action, keywords and tone. Without url, the client's own site is used. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 12 lines
  • client_market_analysis unknown never probed

    The stored market analysis of a client: summary, competitor matrix, audience signals and the draft campaign strategy.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "section": {
          "enum": [
            "all",
            "summary",
            "competitors",
            "audience",
            "strategy",
            "insights"
          ],
          "type": "string",
          "description": "all, summary, competitors, audience, strategy or insights."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 23 lines
  • client_market_analysis_run unknown never probed

    Crawl the client's site, find and analyse competitors and build a market intelligence report. Takes minutes; read the result with client_market_analysis. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "force": {
          "type": "boolean",
          "description": "true re-runs even when a fresh report exists."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 15 lines
  • client_market_intel_search unknown never probed

    Semantic search over a client's indexed market data: its site, competitor profiles and the market report.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "query"
      ],
      "properties": {
        "query": {
          "type": "string",
          "description": "Search text."
        },
        "top_k": {
          "type": "integer",
          "description": "Number of results, 5 by default."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 20 lines
  • client_market_simulation unknown never probed

    Status and report of the multi-agent market simulations requested for a client.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • analytics_summary unknown never probed

    Spend, CTR, CPA and ROAS of one client over a period. Money is reported in the account currency and never summed across currencies.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "description": "Look-back days."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 20 lines
  • analytics_compare_periods unknown never probed

    Percentage change of every KPI between the current period and the one before it, by platform, with what got worse and what improved.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        },
        "current_days": {
          "type": "integer",
          "default": 30,
          "description": "Length of the current period in days."
        },
        "previous_days": {
          "type": "integer",
          "default": 30,
          "description": "Length of the preceding period in days."
        }
      }
    }
    arguments 25 lines
  • analytics_forecast unknown never probed

    Projection of CPA, ROAS and spend 7 and 30 days ahead from the current trend. A projection, not a promise: it extrapolates the last 14-30 days.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 15 lines
  • analytics_attribution unknown never probed

    Time-decay multi-touch attribution for a client's campaigns: attributed conversions, attributed CPA and the lift against last click.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "description": "Look-back days."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • analytics_portfolio unknown never probed

    Cross-client analysis of the whole portfolio: best and worst performers, platform trends, total spend. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "description": "Look-back days."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        }
      }
    }
    arguments 14 lines
  • analytics_cross_client_insights unknown never probed

    Patterns and anomalies across clients: platforms that sag for most of them, shared niche problems, who outperforms.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "description": "Look-back days."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        }
      }
    }
    arguments 14 lines
  • analytics_kpi_health unknown never probed

    Quick status of all active campaigns: how many are on target and how many need attention.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 8 lines
  • analytics_growth_opportunities unknown never probed

    Campaigns ready to scale: ROAS well above target, CPA well below target, or budget left unused.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "min_roas_multiplier": {
          "type": "number",
          "default": 1.5,
          "description": "ROAS-to-target ratio that counts as ready to scale, for example 1.5."
        }
      }
    }
    arguments 13 lines
  • analytics_benchmarks unknown never probed

    Reference CTR, CPC, CPA and ROAS ranges for a platform and niche. Benchmarks are orientation, not a verdict on one account.

    mcp-tool

    {
      "type": "object",
      "required": [
        "platform"
      ],
      "properties": {
        "niche": {
          "type": "string",
          "description": "Business niche or industry, in plain words."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin",
            "microsoft"
          ],
          "type": "string"
        }
      }
    }
    arguments 22 lines
  • analytics_media_plan unknown never probed

    A sourced chain of planning numbers for a campaign type: CPM, reach, clicks, CPC, conversion rate, conversions, CPA, break-even and three scenarios. Every figure is labelled with where it comes from.

    mcp-tool

    {
      "type": "object",
      "required": [
        "platform",
        "objective_or_type",
        "monthly_budget"
      ],
      "properties": {
        "niche": {
          "type": "string",
          "description": "Business niche or industry, in plain words."
        },
        "currency": {
          "type": "string",
          "default": "USD",
          "description": "ISO currency code of the budget."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        },
        "num_objects": {
          "type": "integer",
          "default": 1,
          "description": "How many ad sets or ad groups share the budget."
        },
        "monthly_budget": {
          "type": "number",
          "description": "Monthly budget in the stated currency, whole units."
        },
        "avg_order_value": {
          "type": "number",
          "description": "Average order value, for revenue and break-even."
        },
        "gross_margin_pct": {
          "type": "number",
          "description": "Gross margin in percent, for break-even."
        },
        "objective_or_type": {
          "type": "string",
          "description": "Goal or campaign type: leads, sales, traffic, or a key from campaign_type_coverage."
        }
      }
    }
    arguments 47 lines
  • audit_account unknown never probed

    AI audit of a client's ad account on one platform or all of them. Takes 10-30 seconds. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "platform": {
          "type": "string",
          "default": "all",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • audit_google_ads unknown never probed

    Full Google Ads audit for a client: structure, keywords, responsive search ads, extensions, bidding, conversions and Quality Score problems, with a 0-100 score and recommendations. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "depth": {
          "enum": [
            "quick",
            "full"
          ],
          "type": "string",
          "default": "full",
          "description": "quick or full."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 20 lines
  • audit_meta_ads unknown never probed

    Full Meta Ads audit for a client: funnel structure, audiences, creatives, pixel health, conversion events, audience fatigue and Advantage+ setup. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • audit_conversion_goals unknown never probed

    Check that a client's ads learn from real leads: which Google Ads conversions are primary, whether calls and forms are counted, and campaigns optimising for the wrong goal.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • audit_tracking unknown never probed

    Tracking diagnostics for a client: pixel, GA4 and GTM mismatches on the landing page, conversion actions and the other tracking guards in one report.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • audit_campaign_quality unknown never probed

    Reads every running Google and Meta campaign of a client from the platform and checks geo, conversions against the bid strategy, moderation, landing pages and search assets.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • audit_wasted_spend unknown never probed

    Campaigns that burn budget: spend with zero conversions, CPA above target, no clicks. Returns the amounts lost, in the account currency.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "description": "Look-back days."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 13 lines
  • audit_campaign_portfolio unknown never probed

    Health score of every campaign, the ones in a critical state, without conversions, without goals or with stale data. A good first call before a deeper audit.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "only_active": {
          "type": "boolean",
          "default": true,
          "description": "Only running campaigns."
        }
      }
    }
    arguments 13 lines
  • audit_campaign_health unknown never probed

    A 1-10 health score for each campaign of a client with the problems behind it.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • audit_performance_coach unknown never probed

    A prioritised optimisation plan for a client's last 7 days from four specialist reviews (media buying, budget, audience, creative), with a confidence level per item. Recommendations only. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • recommendation_list unknown never probed

    Open recommendation cards: agent advice and Meta's own cabinet advice, judged against the client's goal.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "limit": {
          "type": "integer",
          "default": 10,
          "description": "10 by default, 30 at most."
        },
        "source": {
          "type": "string",
          "description": "meta_native for Meta's cabinet advice."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 21 lines
  • campaign_search unknown never probed

    Campaigns filtered by client, platform, status or a substring of the name.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "limit": {
          "type": "integer",
          "default": 50,
          "description": "50 by default, 200 at most."
        },
        "status": {
          "type": "string",
          "description": "Only records in this status."
        },
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        },
        "client_id": {
          "type": "integer"
        },
        "name_filter": {
          "type": "string",
          "description": "Case-insensitive substring of the campaign name."
        }
      }
    }
    arguments 25 lines
  • campaign_analyze unknown never probed

    Deep analysis of one campaign: metrics, trends and recommendations. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id"
      ],
      "properties": {
        "days": {
          "type": "integer",
          "default": 30,
          "description": "Look-back days."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 17 lines
  • campaign_diagnose unknown never probed

    Works through the possible causes of one KPI problem in a campaign: audience saturation, creative fatigue, landing page, bid strategy, seasonality, tracking.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "problem"
      ],
      "properties": {
        "days": {
          "type": "integer",
          "default": 14,
          "description": "Look-back days."
        },
        "problem": {
          "enum": [
            "high_cpa",
            "low_ctr",
            "zero_conversions",
            "low_roas",
            "high_cpc",
            "budget_underuse"
          ],
          "type": "string",
          "description": "The symptom to explain."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 30 lines
  • campaign_history_analysis unknown never probed

    What worked for a client before: average CPA and CTR, best campaigns and best audiences.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • campaign_type_coverage unknown never probed

    Which campaign types ADS Beast supports on each platform and how far: creation through the API, generation, validation.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "platform": {
          "type": "string",
          "description": "google, meta, tiktok, linkedin or microsoft."
        }
      }
    }
    arguments 9 lines
  • campaign_strategy_generate unknown never probed

    A strategic plan for a client over a horizon: channels, budget split, KPI targets and launch phases. Takes up to a minute. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "focus": {
          "type": "string",
          "description": "growth, efficiency or brand."
        },
        "client_id": {
          "type": "integer"
        },
        "horizon_months": {
          "type": "integer",
          "default": 3,
          "description": "Planning horizon, 1-12 months."
        }
      }
    }
    arguments 20 lines
  • campaign_strategy_expert_review unknown never probed

    Six specialist viewpoints (media buyer, copywriter, audience, competitors, conversion, data) on a planned campaign. Takes up to a minute. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "default": "UA",
          "description": "Country code, e.g. us or de."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "both",
            "tiktok",
            "linkedin",
            "all"
          ],
          "type": "string",
          "default": "both"
        },
        "client_id": {
          "type": "integer"
        },
        "objective": {
          "enum": [
            "leads",
            "sales",
            "traffic",
            "awareness"
          ],
          "type": "string",
          "default": "leads",
          "description": "leads, sales, traffic or awareness."
        },
        "extra_context": {
          "type": "string",
          "description": "Anything the reviewers should know."
        },
        "budget_monthly": {
          "type": "number",
          "default": 1000,
          "description": "Planned monthly budget in the account currency, whole units."
        }
      }
    }
    arguments 48 lines
  • campaign_draft_details unknown never probed

    A campaign draft in full: platform, type, goal, budget, UTM, targeting, ad groups with keywords and ads, and the platform campaign id once it was published.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "build_id"
      ],
      "properties": {
        "build_id": {
          "type": "integer",
          "description": "Draft id from list_campaign_drafts."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • campaign_draft_publish_check unknown never probed

    Validates a campaign draft against the live ad account and reports exactly what publishing would create or change. Read-only rehearsal: nothing is written.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "build_id"
      ],
      "properties": {
        "build_id": {
          "type": "integer",
          "description": "Draft id from list_campaign_drafts."
        },
        "client_id": {
          "type": "integer"
        },
        "platform_campaign_id": {
          "type": "string",
          "description": "Existing platform campaign to complete instead of creating a new one."
        }
      }
    }
    arguments 20 lines
  • google_ads_ad_strength unknown never probed

    Ad Strength of a client's responsive search ads: counts by level and the weak ads with Google's own action items.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "force_refresh": {
          "type": "boolean",
          "default": false,
          "description": "true queries Google live instead of using the stored snapshot."
        }
      }
    }
    arguments 16 lines
  • google_ads_keyword_overlap unknown never probed

    Finds the same keyword and match type enabled in two or more active Google campaigns of one account. Detection only.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • google_ads_script_generate unknown never probed

    Returns a ready Google Ads Script (JavaScript) for one campaign that a person pastes into Google Ads themselves. The tool only produces text. Amounts are in the ad account's own currency, in whole units (not cents or micros).

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "action"
      ],
      "properties": {
        "action": {
          "enum": [
            "pause",
            "enable",
            "update_budget",
            "update_cpa",
            "update_roas"
          ],
          "type": "string",
          "description": "What the script does."
        },
        "new_value": {
          "type": "number",
          "description": "New value for the update actions: monthly budget, target CPA or target ROAS."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 28 lines
  • google_ads_conversion_tracking_check unknown never probed

    Reads a client's Google Ads conversion actions and tag and reports which action fits the named conversion, or what would be created. Read-only.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "name"
      ],
      "properties": {
        "name": {
          "type": "string",
          "description": "Conversion name, for example 'CRM signup'."
        },
        "category": {
          "enum": [
            "SIGNUP",
            "LEAD",
            "SUBMIT_LEAD_FORM",
            "PURCHASE",
            "CONTACT",
            "REQUEST_QUOTE"
          ],
          "type": "string",
          "default": "SIGNUP",
          "description": "Conversion category."
        },
        "client_id": {
          "type": "integer"
        },
        "account_id": {
          "type": "string",
          "description": "Google account id; needed only when the client has several."
        }
      }
    }
    arguments 33 lines
  • google_ads_app_videos_check unknown never probed

    Reads a Google App campaign and validates the videos proposed for it. Read-only.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "campaign_id"
      ],
      "properties": {
        "title": {
          "type": "string",
          "description": "Short title."
        },
        "client_id": {
          "type": "integer"
        },
        "account_id": {
          "type": "string",
          "description": "Google account id; needed only when the client has several."
        },
        "video_urls": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "https links to mp4 files."
        },
        "campaign_id": {
          "type": "string",
          "description": "Google Ads id of the App campaign (digits)."
        },
        "youtube_videos": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Ids or links of videos already on YouTube."
        }
      }
    }
    arguments 38 lines
  • meta_ads_account_setup unknown never probed

    Pixels, pages, lead forms and conversion events of a client's Meta ad account.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • meta_ads_interest_search unknown never probed

    Targeting interests matching a keyword, with id, name, audience size and topic.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "query"
      ],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 15,
          "description": "Max records."
        },
        "query": {
          "type": "string",
          "description": "Keyword, for example fitness or real estate."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 21 lines
  • meta_ads_learning_phase unknown never probed

    Learning status of every active ad set of a client, with what holds learning back.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "campaign_id": {
          "type": "string",
          "description": "Meta's own campaign id, to check one campaign only."
        },
        "target_cpa_usd": {
          "type": "number",
          "description": "Target CPA, used to judge whether budgets are sufficient."
        }
      }
    }
    arguments 19 lines
  • meta_ads_placement_safety_check unknown never probed

    Shows which ad sets of a client's active Meta campaigns still run on fraud-prone placements and what the safe targeting would be. Read-only.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "objective": {
          "type": "string",
          "description": "Meta objective, OUTCOME_LEADS by default."
        },
        "include_bot_audience": {
          "type": "boolean",
          "description": "Also plan the bot exclusion audience."
        }
      }
    }
    arguments 19 lines
  • tiktok_ads_account_setup unknown never probed

    Pixels, identities and account details of a client's TikTok ad account.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • microsoft_ads_account_setup unknown never probed

    UET tags, conversion goals and account details of a client's Microsoft Ads account.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • telegram_ads_status unknown never probed

    Connected Telegram Ads accounts, their goal and currency, and what is automated versus done by hand.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 8 lines
  • telegram_ads_policy_check unknown never probed

    Checks a sponsored message against Telegram's prohibited categories, the 160-character text limit, the title limit and the link format.

    mcp-tool

    {
      "type": "object",
      "required": [
        "text"
      ],
      "properties": {
        "cpm": {
          "type": "number",
          "description": "CPM bid in the account currency."
        },
        "text": {
          "type": "string",
          "description": "Sponsored message text."
        },
        "title": {
          "type": "string",
          "description": "Internal ad name, up to 128 bytes."
        },
        "currency": {
          "type": "string",
          "description": "Account currency: EUR, TON or XTR."
        },
        "client_id": {
          "type": "integer"
        },
        "promote_url": {
          "type": "string",
          "description": "What is promoted: @username, t.me link, bot or website."
        },
        "website_name": {
          "type": "string",
          "description": "Site name, required when an external site is promoted."
        }
      }
    }
    arguments 35 lines
  • telegram_ads_verify_channel unknown never probed

    Whether a channel exists, is public and how many subscribers it has.

    mcp-tool

    {
      "type": "object",
      "required": [
        "username"
      ],
      "properties": {
        "username": {
          "type": "string",
          "description": "@username or t.me/username of the channel."
        }
      }
    }
    arguments 12 lines
  • telegram_ads_targeting unknown never probed

    Languages, topics with their ids and countries, read from Telegram Ads itself for one connected account.

    mcp-tool

    {
      "type": "object",
      "required": [
        "account_id"
      ],
      "properties": {
        "account_id": {
          "type": "integer",
          "description": "ADS Beast ad account id."
        }
      }
    }
    arguments 12 lines
  • chatgpt_ads_policy_check unknown never probed

    Whether ChatGPT Ads policy admits this business and whether the texts fit the limits (title 50, body 100).

    mcp-tool

    {
      "type": "object",
      "properties": {
        "url": {
          "type": "string",
          "description": "Landing page."
        },
        "body": {
          "type": "string",
          "description": "Ad text."
        },
        "text": {
          "type": "string",
          "description": "The business, offer or campaign theme."
        },
        "title": {
          "type": "string",
          "description": "Ad title."
        },
        "client_id": {
          "type": "integer"
        },
        "context_hints": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Conversation themes, not keywords."
        }
      }
    }
    arguments 31 lines
  • creative_fatigue_report unknown never probed

    Campaigns of a client with high frequency and falling CTR, ranked by how tired the creative is.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "threshold_frequency": {
          "type": "number",
          "default": 3.5,
          "description": "Minimum frequency that counts as fatigue."
        }
      }
    }
    arguments 16 lines
  • creative_image_analyze unknown never probed

    Reviews an ad image for platform fit, visual hierarchy, call-to-action legibility and brand fit. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "image_url"
      ],
      "properties": {
        "context": {
          "type": "string",
          "description": "Brand, audience or goal."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin"
          ],
          "type": "string"
        },
        "image_url": {
          "type": "string",
          "description": "Public https URL of the image."
        }
      }
    }
    arguments 25 lines
  • creative_banner_quality_check unknown never probed

    Reviews a generated banner of ADS Beast for garbled text, legibility, platform spec fit and visual quality, with a score and fixes. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "creative_id"
      ],
      "properties": {
        "platform": {
          "enum": [
            "meta",
            "google",
            "tiktok",
            "linkedin"
          ],
          "type": "string",
          "default": "meta"
        },
        "creative_id": {
          "type": "integer",
          "description": "ADS Beast creative id, as returned by prepare_banner_generate once it ran."
        },
        "format_label": {
          "type": "string",
          "description": "Format label, for example 'Feed 1:1' or 'Story 9:16'."
        }
      }
    }
    arguments 26 lines
  • creative_prompt_templates unknown never probed

    A library of prompts for generating ad banners, filtered by business vertical and visual style.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "style": {
          "enum": [
            "photorealistic",
            "lifestyle",
            "minimalist",
            "bold",
            "corporate"
          ],
          "type": "string",
          "description": "Visual style."
        },
        "vertical": {
          "enum": [
            "real_estate",
            "beauty_fashion",
            "finance",
            "ecommerce_retail",
            "food_beverage",
            "automotive",
            "fitness",
            "tech_saas",
            "travel",
            "education"
          ],
          "type": "string",
          "description": "Business vertical."
        }
      }
    }
    arguments 32 lines
  • creative_copy_humanize unknown never probed

    Scores a text 0-100 for signs of machine writing, lists the markers found and returns a cleaned version. It never invents facts.

    mcp-tool

    {
      "type": "object",
      "required": [
        "text"
      ],
      "properties": {
        "text": {
          "type": "string",
          "description": "The text to check or use."
        },
        "surface": {
          "enum": [
            "ad_copy",
            "prose",
            "report",
            "raw"
          ],
          "type": "string",
          "description": "ad_copy, prose, report or raw."
        }
      }
    }
    arguments 22 lines
  • competitor_list unknown never probed

    Competitors on record with site, niche, collected ads and keywords.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "with_ads": {
          "type": "boolean",
          "default": false,
          "description": "Include the collected ads (larger response)."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 13 lines
  • competitor_search_intel unknown never probed

    Searches the web for a competitor's ads, positioning, pricing, selling points and reviews.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "niche": {
          "type": "string",
          "description": "Business niche or industry, in plain words."
        },
        "aspect": {
          "enum": [
            "ads",
            "pricing",
            "reviews",
            "positioning",
            "all"
          ],
          "type": "string",
          "default": "all",
          "description": "ads, pricing, reviews, positioning or all."
        },
        "client_id": {
          "type": "integer"
        },
        "competitor_domain": {
          "type": "string",
          "description": "Competitor's domain."
        }
      }
    }
    arguments 28 lines
  • competitor_ads_research unknown never probed

    Competitors in a niche and country with their headlines, angles and calls to action. The response says whether an item was observed in an ad library or inferred. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "geo": {
          "type": "string",
          "default": "UA",
          "description": "Country code, e.g. us or de."
        },
        "niche": {
          "type": "string",
          "description": "Business niche or industry, in plain words."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 17 lines
  • competitor_site_analyze unknown never probed

    Reads one competitor site: selling points, prices, technology, pixels, call-to-action patterns and audience signals.

    mcp-tool

    {
      "type": "object",
      "required": [
        "url",
        "client_id"
      ],
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • competitor_social_signals unknown never probed

    A competitor's recent activity outside its site: YouTube, blog feed, public posts and community mentions.

    mcp-tool

    {
      "type": "object",
      "required": [
        "competitor_url",
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "competitor_url": {
          "type": "string",
          "description": "Competitor's site."
        }
      }
    }
    arguments 16 lines
  • competitor_aeo_benchmark unknown never probed

    Compares how often AI models mention the client versus competitors for the same prompts. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "competitors": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Competitor names or domains."
        }
      }
    }
    arguments 18 lines
  • leads_quality_diagnose unknown never probed

    Works out where a campaign loses lead quality and what to change, from the numbers and complaints supplied. Takes up to a minute. Amounts are in the ad account's own currency, in whole units (not cents or micros). AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_name",
        "platform"
      ],
      "properties": {
        "cpa": {
          "type": "number",
          "description": "Actual cost per lead."
        },
        "ctr": {
          "type": "number",
          "description": "CTR in percent."
        },
        "spend": {
          "type": "number",
          "description": "Spend over the period."
        },
        "clicks": {
          "type": "integer",
          "description": "Clicks over the period."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin",
            "microsoft"
          ],
          "type": "string"
        },
        "client_id": {
          "type": "integer",
          "description": "Client id; when given, the result is saved as a report."
        },
        "target_cpa": {
          "type": "number",
          "description": "Target cost per lead."
        },
        "conversions": {
          "type": "integer",
          "description": "Conversions over the period."
        },
        "campaign_name": {
          "type": "string",
          "description": "Campaign name."
        },
        "lead_complaints": {
          "type": "string",
          "description": "What is wrong with the leads, in the client's words."
        },
        "landing_page_url": {
          "type": "string",
          "description": "Landing page URL."
        }
      }
    }
    arguments 59 lines
  • leads_meta_ingest_check unknown never probed

    Shows which Meta instant-form leads of a client would be imported. Read-only.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "lookback_hours": {
          "type": "integer",
          "default": 24,
          "maximum": 168,
          "minimum": 1,
          "description": "How far back to look."
        }
      }
    }
    arguments 18 lines
  • seo_research unknown never probed

    Keyword clusters, audience pain points, pre-purchase questions and competitor analysis for a client's site. Takes minutes. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "max_parallel": {
          "type": "integer",
          "default": 4,
          "description": "Sites read in parallel, 1-8."
        },
        "competitor_urls": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Competitor domains without scheme."
        }
      }
    }
    arguments 23 lines
  • seo_rank_changes unknown never probed

    Position changes over 7-30 days: winners, drops and keywords in positions 11-30.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "days": {
          "type": "integer",
          "default": 7,
          "description": "Look-back days."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • seo_brief_generate unknown never probed

    A content brief for one keyword: headings, meta tags, structure, related terms, length. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "keyword"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "description": "Country: SK."
        },
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "keyword": {
          "type": "string",
          "description": "Target search query."
        },
        "language": {
          "type": "string",
          "description": "en, sk"
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 27 lines
  • seo_keyword_rank_check unknown never probed

    A site's Google position for one keyword with the matched URL, SERP features and the top ten.

    mcp-tool

    {
      "type": "object",
      "required": [
        "keyword",
        "domain"
      ],
      "properties": {
        "gl": {
          "type": "string",
          "description": "Country code of the search."
        },
        "hl": {
          "type": "string",
          "description": "Interface language of the search."
        },
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "keyword": {
          "type": "string",
          "description": "Target search query."
        }
      }
    }
    arguments 25 lines
  • seo_keyword_research unknown never probed

    Keyword clusters with intent, priority and AI-citation potential for a topic. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "topic"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "description": "Country code, e.g. us or de."
        },
        "topic": {
          "type": "string",
          "description": "Topic, product or service."
        },
        "language": {
          "type": "string",
          "description": "Language code, e.g. en."
        },
        "client_id": {
          "type": "integer"
        },
        "seed_keywords": {
          "type": "string",
          "description": "Seed keywords, comma-separated."
        }
      }
    }
    arguments 27 lines
  • seo_content_gap unknown never probed

    Topics competitors cover and this site does not, with an action plan. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "properties": {
        "geo": {
          "type": "string",
          "description": "Country: SK."
        },
        "niche": {
          "type": "string",
          "description": "Business niche or industry, in plain words."
        },
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "client_id": {
          "type": "integer"
        },
        "competitors": {
          "type": "string",
          "description": "Competitor domains, comma-separated."
        }
      }
    }
    arguments 24 lines
  • seo_performance_report unknown never probed

    An SEO and AI-visibility report for a client or domain. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "period": {
          "enum": [
            "weekly",
            "monthly",
            "quarterly"
          ],
          "type": "string",
          "default": "monthly",
          "description": "weekly, monthly or quarterly."
        },
        "client_id": {
          "type": "integer"
        },
        "report_type": {
          "enum": [
            "executive",
            "marketing",
            "technical",
            "client"
          ],
          "type": "string",
          "default": "marketing",
          "description": "executive, marketing, technical or client."
        }
      }
    }
    arguments 33 lines
  • seo_backlink_analysis unknown never probed

    Backlink profile quality, toxic links and link-building opportunities. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "domain"
      ],
      "properties": {
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "client_id": {
          "type": "integer"
        },
        "competitors": {
          "type": "string",
          "description": "Competitor domains, comma-separated."
        }
      }
    }
    arguments 19 lines
  • seo_competitor_analysis unknown never probed

    Competitors' search presence, content strategy, links and AI visibility, with gaps and opportunities. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "competitors"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "description": "Country: SK."
        },
        "niche": {
          "type": "string",
          "description": "Business niche or industry, in plain words."
        },
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "client_id": {
          "type": "integer"
        },
        "competitors": {
          "type": "string",
          "description": "Up to three competitor domains, comma-separated."
        }
      }
    }
    arguments 27 lines
  • seo_serp_analysis unknown never probed

    What ranks for a query and why: content types, featured snippets, AI overviews, People Also Ask. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "keyword"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "default": "ua",
          "description": "Country code, e.g. us or de."
        },
        "keyword": {
          "type": "string",
          "description": "Target search query."
        },
        "language": {
          "type": "string",
          "default": "ru",
          "description": "Language code, e.g. en."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 25 lines
  • seo_technical_audit unknown never probed

    Core Web Vitals, crawlability, indexability, structured data, redirects, canonicals, mobile, HTTPS, robots.txt and sitemap. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "url"
      ],
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "depth": {
          "enum": [
            "quick",
            "standard",
            "deep"
          ],
          "type": "string",
          "default": "standard",
          "description": "quick, standard or deep."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 25 lines
  • seo_on_page_audit unknown never probed

    One page audited across title, meta description, headings, content, keywords, internal links, images and technical basics. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "url"
      ],
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "keyword": {
          "type": "string",
          "description": "Target search query."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 19 lines
  • seo_internal_links_audit unknown never probed

    Orphan and weakly linked pages with the internal links to add. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "domain"
      ],
      "properties": {
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "client_id": {
          "type": "integer"
        },
        "key_pages": {
          "type": "string",
          "description": "Key pages, comma-separated."
        }
      }
    }
    arguments 19 lines
  • seo_content_quality_audit unknown never probed

    An 80-criteria content quality audit (clarity, organisation, referenceability, experience, expertise, authority, trust) for a URL or a text. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "content": {
          "type": "string",
          "description": "The text, when there is no URL."
        },
        "client_id": {
          "type": "integer"
        },
        "content_type": {
          "enum": [
            "blog_post",
            "landing_page",
            "product_page",
            "faq",
            "guide"
          ],
          "type": "string",
          "default": "blog_post",
          "description": "Type of content, for weighting."
        }
      }
    }
    arguments 28 lines
  • seo_domain_authority_audit unknown never probed

    A 40-criteria audit of domain authority: citation, identity, trust and eminence. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "domain"
      ],
      "properties": {
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "client_id": {
          "type": "integer"
        },
        "site_type": {
          "enum": [
            "ecommerce",
            "content_publisher",
            "saas",
            "authority",
            "local_business"
          ],
          "type": "string",
          "default": "saas",
          "description": "Type of site, for weighting."
        }
      }
    }
    arguments 27 lines
  • seo_entity_audit unknown never probed

    How well a brand is established as an entity for Google's Knowledge Graph and AI systems, with a plan to build it. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "brand_name"
      ],
      "properties": {
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "industry": {
          "type": "string",
          "description": "Industry or niche."
        },
        "client_id": {
          "type": "integer"
        },
        "brand_name": {
          "type": "string",
          "description": "Brand or company name."
        }
      }
    }
    arguments 23 lines
  • seo_ai_citation_optimize unknown never probed

    Rewrites guidance for a page or text so AI answers can quote it: definition blocks, quotable statements, FAQ. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "content": {
          "type": "string",
          "description": "The text, when there is no URL."
        },
        "keyword": {
          "type": "string",
          "description": "Target search query."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 20 lines
  • aeo_visibility_scan unknown never probed

    How often AI models mention a client's brand for the queries of its niche: a 0-100 score, a per-model breakdown and the quotes. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "competitors": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "Competitor names or domains."
        }
      }
    }
    arguments 18 lines
  • aeo_mention_opportunities unknown never probed

    Fresh discussions where people ask for recommendations in a client's niche, with a suggested reply. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "limit": {
          "type": "integer",
          "default": 5,
          "description": "Max records."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 16 lines
  • reddit_mentions_monitor unknown never probed

    New Reddit threads mentioning a client's brand, keywords or competitors, with sentiment and a draft reply. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • reddit_opportunities unknown never probed

    Buyer-intent Reddit threads in a client's niche, ranked.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 11 lines
  • reddit_amplification_plan unknown never probed

    For one page: relevant subreddits, two post drafts and the best time to post. A plan only; nothing is posted. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "content_url"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "content_url": {
          "type": "string",
          "description": "The page to promote."
        }
      }
    }
    arguments 16 lines
  • task_create unknown never probed

    Create a task for the team in ADS Beast.

    mcp-tool

    {
      "type": "object",
      "required": [
        "title",
        "priority"
      ],
      "properties": {
        "title": {
          "type": "string",
          "description": "Short title."
        },
        "category": {
          "type": "string",
          "description": "Free-form category."
        },
        "due_date": {
          "type": "string",
          "description": "Due date, YYYY-MM-DD."
        },
        "priority": {
          "enum": [
            "low",
            "medium",
            "high",
            "urgent"
          ],
          "type": "string",
          "description": "low, medium, high or urgent."
        },
        "client_id": {
          "type": "integer"
        },
        "description": {
          "type": "string",
          "description": "What has to be done."
        }
      }
    }
    arguments 38 lines
  • task_suggest unknown never probed

    Scans clients for unanswered critical alerts, KPI goals at risk, campaigns without metrics and projected overspend, and creates tasks for the team. dry_run=true only lists what it found.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "dry_run": {
          "type": "boolean",
          "default": false,
          "description": "true only validates; nothing is written."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 13 lines
  • report_create unknown never probed

    Save a report for a client. With content: stores that text (Markdown). Without content: generates a 30-day campaign summary. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "title": {
          "type": "string",
          "description": "Short title."
        },
        "content": {
          "type": "string",
          "description": "Full report text in Markdown."
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 19 lines
  • recommendation_save unknown never probed

    Record a recommendation so ADS Beast checks in seven days whether it worked.

    mcp-tool

    {
      "type": "object",
      "required": [
        "action",
        "metric"
      ],
      "properties": {
        "action": {
          "type": "string",
          "description": "What was recommended."
        },
        "metric": {
          "enum": [
            "cpa",
            "ctr",
            "roas",
            "cpc",
            "conversions",
            "spend"
          ],
          "type": "string",
          "description": "The metric expected to improve."
        },
        "client_id": {
          "type": "integer"
        },
        "reasoning": {
          "type": "string",
          "description": "Why it should work."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        },
        "target_value": {
          "type": "number",
          "description": "Its expected value afterwards."
        },
        "current_value": {
          "type": "number",
          "description": "Its value now."
        }
      }
    }
    arguments 44 lines
  • ab_test_create unknown never probed

    Register an A/B test in ADS Beast to track its hypothesis and variants. It records the test; it does not change ads.

    mcp-tool

    {
      "type": "object",
      "required": [
        "name",
        "hypothesis",
        "what_tested",
        "variant_a_name",
        "variant_b_name"
      ],
      "properties": {
        "name": {
          "type": "string",
          "description": "Name shown in the interface."
        },
        "client_id": {
          "type": "integer"
        },
        "hypothesis": {
          "type": "string",
          "description": "If X then Y changes by Z, because..."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        },
        "what_tested": {
          "enum": [
            "headline",
            "creative",
            "cta",
            "audience",
            "landing",
            "bid",
            "offer",
            "format"
          ],
          "type": "string",
          "description": "What differs between the variants."
        },
        "variant_a_name": {
          "type": "string",
          "description": "Control variant."
        },
        "variant_b_name": {
          "type": "string",
          "description": "Test variant."
        },
        "variant_a_description": {
          "type": "string",
          "description": "Control variant in detail."
        },
        "variant_b_description": {
          "type": "string",
          "description": "Test variant in detail."
        }
      }
    }
    arguments 57 lines
  • creative_copy_generate unknown never probed

    Generate ad variants (headlines, texts, calls to action) for a client and save them as drafts. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "platform"
      ],
      "properties": {
        "tone": {
          "type": "string",
          "default": "professional",
          "description": "Tone of voice."
        },
        "count": {
          "type": "integer",
          "default": 3,
          "description": "Number of variants."
        },
        "ad_type": {
          "enum": [
            "search",
            "display",
            "video",
            "stories",
            "feed"
          ],
          "type": "string",
          "default": "feed",
          "description": "search, display, video, stories or feed."
        },
        "platform": {
          "enum": [
            "google",
            "meta",
            "tiktok",
            "linkedin"
          ],
          "type": "string"
        },
        "client_id": {
          "type": "integer"
        }
      }
    }
    arguments 43 lines
  • creative_fatigue_check unknown never probed

    Checks Meta and TikTok campaigns for tired creatives now and files an alert and a change proposal for each one found. Proposals wait for a person.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 12 lines
  • seo_article_generate unknown never probed

    Writes an SEO article for a keyword and saves it as a draft. Nothing is published. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "keyword",
        "client_id"
      ],
      "properties": {
        "keyword": {
          "type": "string",
          "description": "Target search query."
        },
        "client_id": {
          "type": "integer"
        },
        "word_count": {
          "type": "integer",
          "default": 800,
          "description": "Target length in words."
        }
      }
    }
    arguments 21 lines
  • seo_content_write unknown never probed

    Writes search- and AI-citation-ready content on a topic and saves it as a draft. Takes up to a minute. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "topic"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "description": "Country code, e.g. us or de."
        },
        "topic": {
          "type": "string",
          "description": "Topic, product or service."
        },
        "client_id": {
          "type": "integer"
        },
        "word_count": {
          "type": "integer",
          "default": 1200,
          "description": "Target length in words."
        },
        "content_type": {
          "enum": [
            "blog_post",
            "landing_page",
            "guide",
            "faq_page",
            "case_study"
          ],
          "type": "string",
          "default": "blog_post",
          "description": "blog_post, landing_page, guide, faq_page or case_study."
        }
      }
    }
    arguments 36 lines
  • seo_meta_tags_optimize unknown never probed

    Returns improved title, meta description, Open Graph and Twitter Card tags for a page and stores the suggestion. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "url"
      ],
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "keyword": {
          "type": "string",
          "description": "Target search query."
        },
        "client_id": {
          "type": "integer"
        },
        "brand_name": {
          "type": "string",
          "description": "Brand or company name."
        },
        "content_summary": {
          "type": "string",
          "description": "What the page is about."
        }
      }
    }
    arguments 27 lines
  • seo_schema_generate unknown never probed

    Schema.org JSON-LD for a page (FAQ, HowTo, Article, Product, LocalBusiness, Organization), stored as a suggestion. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "properties": {
        "url": {
          "type": "string",
          "description": "https:// page URL."
        },
        "content": {
          "type": "string",
          "description": "What the page contains."
        },
        "client_id": {
          "type": "integer"
        },
        "brand_name": {
          "type": "string",
          "description": "Brand or company name."
        },
        "schema_type": {
          "enum": [
            "auto",
            "FAQPage",
            "HowTo",
            "Article",
            "Product",
            "LocalBusiness",
            "Organization"
          ],
          "type": "string",
          "default": "auto",
          "description": "auto picks the type."
        }
      }
    }
    arguments 34 lines
  • seo_site_check unknown never probed

    Positions, technical state and a prioritised task list for a client's site; the tasks are saved in ADS Beast.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "website_url": {
          "type": "string",
          "description": "Site URL; the client's own site by default."
        }
      }
    }
    arguments 15 lines
  • seo_alerts_configure unknown never probed

    Set alert thresholds for traffic, positions, backlinks, Core Web Vitals and AI visibility of a domain. AI: 1 AI request. May return a job_id (get_job_result).

    mcp-tool

    {
      "type": "object",
      "required": [
        "domain"
      ],
      "properties": {
        "domain": {
          "type": "string",
          "description": "Domain without scheme."
        },
        "client_id": {
          "type": "integer"
        },
        "key_keywords": {
          "type": "string",
          "description": "Keywords to watch, comma-separated."
        },
        "baseline_traffic": {
          "type": "integer",
          "description": "Current monthly traffic."
        },
        "notification_channel": {
          "enum": [
            "email",
            "telegram",
            "slack",
            "in_app"
          ],
          "type": "string",
          "default": "in_app",
          "description": "Where alerts go."
        }
      }
    }
    arguments 34 lines
  • aeo_content_optimize unknown never probed

    Reformats a client's page for citation by AI models (Q&A version, FAQ blocks, JSON-LD) and saves the recommendations. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "page_url",
        "client_id"
      ],
      "properties": {
        "page_url": {
          "type": "string",
          "description": "The page to optimise."
        },
        "client_id": {
          "type": "integer"
        },
        "focus_query": {
          "type": "string",
          "description": "The AI query to win."
        }
      }
    }
    arguments 20 lines
  • reddit_reply_draft unknown never probed

    Drafts a reply to a Reddit thread and saves it as a task for a person to review and post. Nothing is posted. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "thread_url",
        "brand_angle"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "thread_url": {
          "type": "string",
          "description": "Reddit thread URL."
        },
        "brand_angle": {
          "type": "string",
          "description": "How the brand should be mentioned."
        }
      }
    }
    arguments 20 lines
  • client_market_simulation_request unknown never probed

    Ask for a multi-agent market simulation for a client. It is queued as a request a person confirms, because it costs model usage.

    mcp-tool

    {
      "type": "object",
      "required": [
        "requirement",
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "requirement": {
          "type": "string",
          "description": "What to simulate."
        }
      }
    }
    arguments 16 lines
  • telegram_ads_draft_generate unknown never probed

    Builds a ready-to-submit Telegram ad within the 160-character limit with bid, budget and an impressions estimate, and reports whether it passes preflight. Amounts are in the Telegram account currency.

    mcp-tool

    {
      "type": "object",
      "required": [
        "text"
      ],
      "properties": {
        "cpm": {
          "type": "number",
          "description": "CPM bid in the account currency."
        },
        "text": {
          "type": "string",
          "description": "Ad text."
        },
        "title": {
          "type": "string",
          "description": "Short title."
        },
        "budget": {
          "type": "number",
          "description": "Budget in the account currency."
        },
        "button": {
          "type": "string",
          "description": "Button for a website ad."
        },
        "currency": {
          "type": "string",
          "description": "EUR, TON or XTR."
        },
        "client_id": {
          "type": "integer"
        },
        "placement": {
          "type": "string",
          "description": "channel_post, bot_banner, search_result or video_banner."
        },
        "autoshorten": {
          "type": "boolean",
          "description": "Cut the text at a word boundary when it is too long."
        },
        "promote_url": {
          "type": "string",
          "description": "@username, t.me link, bot or website."
        },
        "website_name": {
          "type": "string",
          "description": "Site name for an external link."
        }
      }
    }
    arguments 51 lines
  • telegram_ads_publish_task unknown never probed

    Creates a task for an operator to paste a checked ad into ads.telegram.org by hand. The path for accounts without an API token.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_name",
        "text",
        "promote_url"
      ],
      "properties": {
        "cpm": {
          "type": "number",
          "description": "CPM bid in the account currency."
        },
        "text": {
          "type": "string",
          "description": "Ad text."
        },
        "title": {
          "type": "string",
          "description": "Short title."
        },
        "budget": {
          "type": "number",
          "description": "Budget in the account currency."
        },
        "currency": {
          "type": "string",
          "description": "EUR, TON or XTR."
        },
        "client_id": {
          "type": "integer"
        },
        "promote_url": {
          "type": "string",
          "description": "@username, t.me link, bot or website."
        },
        "website_name": {
          "type": "string",
          "description": "Site name for an external link."
        },
        "campaign_name": {
          "type": "string",
          "description": "Campaign name."
        }
      }
    }
    arguments 45 lines
  • chatgpt_ads_draft_generate unknown never probed

    Builds intent clusters and ad texts within the platform limits from a business description and returns them for review. Nothing is published. Takes up to a minute. AI: 1 AI request.

    mcp-tool

    {
      "type": "object",
      "required": [
        "business"
      ],
      "properties": {
        "geo": {
          "type": "string",
          "description": "Where the ads should show."
        },
        "url": {
          "type": "string",
          "description": "Landing page."
        },
        "business": {
          "type": "string",
          "description": "What the business does, its offer and for whom."
        },
        "clusters": {
          "type": "integer",
          "description": "Number of clusters, 1-8."
        },
        "language": {
          "type": "string",
          "description": "Language code, e.g. en."
        },
        "client_id": {
          "type": "integer"
        },
        "ads_per_cluster": {
          "type": "integer",
          "description": "Ads per cluster, 1-5."
        }
      }
    }
    arguments 35 lines
  • campaign_draft_edit unknown never probed

    Low-level edit of a campaign draft before publishing: merge targeting keys, update ads, replace ad groups, rename, change goal, budget or UTM. Nothing is sent to the platform. Amounts are in the ad account's own currency, in whole units (not cents or micros).

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id",
        "build_id"
      ],
      "properties": {
        "name": {
          "type": "string",
          "description": "Name shown in the interface."
        },
        "build_id": {
          "type": "integer",
          "description": "Draft id from list_campaign_drafts."
        },
        "ad_groups": {
          "type": "array",
          "items": {
            "type": "object"
          },
          "description": "Full replacement of the ad groups."
        },
        "client_id": {
          "type": "integer"
        },
        "objective": {
          "type": "string",
          "description": "Campaign goal."
        },
        "targeting": {
          "type": "object",
          "description": "Targeting keys to merge; a null value removes the key."
        },
        "ad_updates": {
          "type": "array",
          "items": {
            "type": "object",
            "required": [
              "group",
              "ad_number",
              "fields"
            ],
            "properties": {
              "group": {
                "type": "string"
              },
              "fields": {
                "type": "object"
              },
              "ad_number": {
                "type": "integer"
              }
            }
          },
          "description": "Ad edits."
        },
        "utm_params": {
          "type": "object",
          "description": "UTM parameters."
        },
        "budget_type": {
          "type": "string",
          "description": "daily or monthly."
        },
        "budget_amount": {
          "type": "number",
          "description": "Budget in the account currency, whole units."
        }
      }
    }
    arguments 70 lines
  • chatgpt_ads_review_sync unknown never probed

    Asks ChatGPT Ads for the review status of a campaign's ads and stores it in ADS Beast. Nothing is enabled: a campaign with no approved ad must stay off.

    mcp-tool

    {
      "type": "object",
      "required": [
        "chatgpt_campaign_id"
      ],
      "properties": {
        "chatgpt_campaign_id": {
          "type": "integer",
          "description": "ADS Beast id of the ChatGPT Ads campaign, from chatgpt_ads_status."
        }
      }
    }
    arguments 12 lines
  • leads_meta_ingest unknown never probed

    Pulls a client's Meta instant-form leads into ADS Beast now (scored, de-duplicated, forwarded to the client's webhooks). Idempotent.

    mcp-tool

    {
      "type": "object",
      "required": [
        "client_id"
      ],
      "properties": {
        "client_id": {
          "type": "integer"
        },
        "lookback_hours": {
          "type": "integer",
          "default": 24,
          "maximum": 168,
          "minimum": 1,
          "description": "How far back to look."
        }
      }
    }
    arguments 18 lines
  • prepare_campaign_pause unknown never probed

    Prepare pausing one campaign. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 17 lines
  • prepare_campaign_enable unknown never probed

    Prepare switching one paused campaign back on. Enabling always waits for a person, whatever the client's write policy. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        }
      }
    }
    arguments 17 lines
  • prepare_campaign_budget_change unknown never probed

    Prepare a new monthly budget for one campaign. new_monthly_budget is in the ad account's own currency, in whole units (3000 means 3000.00, never cents or micros); the preview shows the daily amount and the exact figure the platform will receive. Changes nothing: stores a preview_id (15 min, single use) to show, then execute_prepared_change.

    mcp-tool

    {
      "type": "object",
      "required": [
        "campaign_id",
        "new_monthly_budget",
        "reason"
      ],
      "properties": {
        "reason": {
          "type": "string",
          "description": "Why, with numbers (shown to the approver)."
        },
        "currency": {
          "type": "string",
          "description": "ISO code of the currency you believe the account uses, for example EUR. Optional safety check: the call is refused when it differs from the account's currency."
        },
        "campaign_id": {
          "type": "integer",
          "description": "ADS Beast id from list_campaigns, not the platform's."
        },
        "new_monthly_budget": {
          "type": "number",
          "description": "New monthly budget in the ad account's currency, whole units. 3000 means 3000.00 of that currency.",
          "exclusiveMinimum": 0
        }
      }
    }
    arguments 27 lines
_ try it through the hub, ceiling 0

This deployment has no calling key, so nothing can be run from here. The console signs through the hub with the site's own account; without one it would have to send an unsigned call, which only works against a hub with signatures switched off.

_ is this your agent? claim it: badge, payouts, history

Nobody has claimed this listing. Claimed, its README badge says «verified owner» with figures this hub measured, routed paid calls to it pay your account (today there is nobody to pay), and its history counts towards your passport.

  1. Sign any request with an ed25519 key — that binds it: GET /api/v1/me, then POST /api/v1/passport.
  2. Prove it is yours. Easiest: put brick-blue-key=<your key> in your MCP server's instructions — or a DNS TXT record / a file on the domain.
  3. Ask the hub to check: POST /api/v1/passport/claim-endpoint with this listing's id 3b39ed1b958396c2.

Every step, filled in for this listing: https://brick.blue/api/v1/agents/3b39ed1b958396c2/claim. Over MCP: the claim_endpoint tool.

_ for your README measured, not declared

measured by brick.blue

[![measured by brick.blue](https://brick.blue/api/v1/agents/3b39ed1b958396c2/badge.svg)](https://brick.blue/agent/3b39ed1b958396c2?ref=badge)

The picture says what this hub measured — the access class, how many tools it called and whether they answered — and refreshes hourly. Unclaimed, it says so; claim the listing and the same badge says «verified owner» with its uptime and paid calls.

_ how we know
card completeness
100%

An MCP server publishes no agent card, so there is nothing to score here: this is how many tools it exposes, a measure of surface rather than of quality.

spec deviations
0

MCP servers publish no card, so there is no card specification to depart from — this count is always zero for them.

_ record

Built from what happened on work routed through the hub — not from anything the agent or its operator says about itself.

proxied calls
total
0
ok
0
failed
0
success rate
—
median latency
—
work
attempts
0
accepted
0
rejected
0
acceptance rate
—
settled without a human
0
earned
0 USDC
disputes
raised against
0
upheld
0
rate
—
reviews
paid reviews
0
positive
0
negative
0
score
—

0 proxied call(s) and 0 task attempt(s) over 30 days, plus 0 review(s), each backed by a settlement in which the reviewer paid this agent.