Tuski

endpoint

https://tuski.pay.zeroclick.io/mcp

protocol

http-sse ·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 Tuski live?

Yes — it answered the hub's last check (checked 1h ago). It answered 99% of checks over the last 30 days.

Is Tuski free to use?

Not anonymously — it asks for a key or a login before it will serve. Whether that key costs anything is the operator's plan to say: the hub measured the door, not the pricing.

What tools does Tuski have?

27 tools: get_company_search, get_contacts, get_job_status, get_enrich_list, unlock_contacts, get_results, get_results_csv, estimate_enrich_list, ….

Is Tuski 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.

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_company_search reads auth-required 9h ago

  Get one company search's status, progress and billing. `search_id` is the id returned by `start_company_search`. Returns the same projection `start_company_search` does, plus `done` (true once the search reached a terminal state: succeeded, failed, or canceled) and a `next_step` that depends on how it ended. `billing` counts `billable_results` (charged) and `free_results` (delivered free, headcount not stated); `effective_rules.company_filters` lists the verified filters, `warnings` what they cannot check, and `diagnostics.blocking_filters` the filters that stopped most candidates. A search that does not exist, or belongs to another organization, raises the same `search_not_found` error either way. Requires the `discovery` permission.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "get_company_searchArguments",
    "required": [
      "search_id"
    ],
    "properties": {
      "search_id": {
        "type": "integer",
        "title": "Search Id"
      }
    }
  }

  arguments 13 lines copy
• get_contacts reads auth-required 9h ago

  List one page of the people found for a contacts job, paginated. `job_id` is the id `find_contacts` returned. `limit` (default 25, clamped into 1..100) and `offset` (default 0, floored at 0) page through the results; `domain` narrows the page to one company. Each row shows email/linkedin/phone only for fields already unlocked for that person; a field never unlocked (or still being unlocked) is always empty here, never the real value ahead of payment. Call `estimate_contact_unlock` and `unlock_contacts` to reveal more fields, then call this again to see them. Returns `{rows, total, limit, offset, has_more, next_offset}`. Requires the `contacts` permission.

  mcp-tool

  1234567891011121314151617181920212223242526272829303132333435

  {
    "type": "object",
    "title": "get_contactsArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "limit": {
        "type": "integer",
        "title": "Limit",
        "default": 25
      },
      "domain": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Domain",
        "default": null
      },
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      },
      "offset": {
        "type": "integer",
        "title": "Offset",
        "default": 0
      }
    }
  }

  arguments 35 lines copy
• get_job_status reads auth-required 9h ago

  Get one job's status, counts and ETA. Returns status, total, queued, sent, valid, invalid, unknown, unreachable, mailbox_full, blocked, soft_temp, routing_error, other_bounce and eta. `unreachable` is the no-MX bucket that gets refunded when the job closes. Call wait_for_job instead if you want to block until the job finishes. No permission required.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "get_job_statusArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      }
    }
  }

  arguments 13 lines copy
• get_enrich_list auth-required 12h ago

  Get a person list's status and counts with a page of its rows. Requires `contacts`. Returns status (running, completed, cancelled), total, done, found, not_found, error, skipped, queued, credits_spent, reused, `finished` and `poll_after_seconds` (wait that long, then call again), plus `rows` (`limit` default 50, at most 100; `offset` pages on), count, total_rows and has_more. A row has status (found, not_found, queued, running, skipped, error), reason, credits_charged and, for a person found, `result` (email, phone, linkedin_url, title, company ...). A list that ended early keeps what it found; skipped rows were never charged.

  mcp-tool

  1234567891011121314151617181920212223

  {
    "type": "object",
    "title": "get_enrich_listArguments",
    "required": [
      "list_id"
    ],
    "properties": {
      "limit": {
        "type": "integer",
        "title": "Limit",
        "default": 50
      },
      "offset": {
        "type": "integer",
        "title": "Offset",
        "default": 0
      },
      "list_id": {
        "type": "integer",
        "title": "List Id"
      }
    }
  }

  arguments 23 lines copy
• unlock_contacts moves money auth-required never probed

  Reveal contact fields for specific people. Spends credits per field, up front. Call `estimate_contact_unlock` first and show the user the price. `selections`: list of `{person_id, fields}`, fields "email", "linkedin", "phone". Owned fields are free, undeliverable ones refunded, a guessed or personal email costs `email_unverified_price`. `verify=true` also probes each unlocked guessed or unknown email (one validation job, `verify_price_per_email` each); its `email_quality` then reads `verified_by_probe` or the failure. The same `idempotency_key` and request never charges twice; without one a key is derived for 10 minutes ("auto:" keys are rejected). 10 calls a minute. Returns `request_ids`, `status`, `balance`, `verify`, `idempotency`, `idempotent_hit`. Call `wait_for_contacts` next. Requires `contacts`.

  mcp-tool

  123456789101112131415161718192021222324252627282930313233343536373839

  {
    "type": "object",
    "title": "unlock_contactsArguments",
    "required": [
      "job_id",
      "selections"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      },
      "verify": {
        "type": "boolean",
        "title": "Verify",
        "default": false
      },
      "selections": {
        "type": "array",
        "items": {
          "type": "object",
          "additionalProperties": true
        },
        "title": "Selections"
      },
      "idempotency_key": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Idempotency Key",
        "default": null
      }
    }
  }

  arguments 39 lines copy
• get_results reads auth-required 7h ago

  Return job results as structured JSON rows (default) or CSV text. Requires `export`. format="json": {job_id, rows, count, limit, offset, has_more}; `limit` defaults to 100, clamped into 1..1000; a negative `offset` is floored at 0. The response echoes the limit/offset actually used. format="csv": {job_id, csv, row_count, has_more} up to 500 rows; above that, {job_id, download_url, expires_in_seconds, row_count, has_more, next_step} instead: a signed link good for a few minutes, no API key needed. Leaving `limit` unset means everything up to the 10,000-row/5MB ceiling, same as get_results_csv. `status_filter` restricts rows to one leads.status value; an unknown value is rejected with the accepted list.

  mcp-tool

  1234567891011121314151617181920212223242526272829303132333435363738394041424344454647

  {
    "type": "object",
    "title": "get_resultsArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "limit": {
        "anyOf": [
          {
            "type": "integer"
          },
          {
            "type": "null"
          }
        ],
        "title": "Limit",
        "default": null
      },
      "format": {
        "type": "string",
        "title": "Format",
        "default": "json"
      },
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      },
      "offset": {
        "type": "integer",
        "title": "Offset",
        "default": 0
      },
      "status_filter": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Status Filter",
        "default": null
      }
    }
  }

  arguments 47 lines copy
• get_results_csv reads auth-required 7h ago

  Return the results CSV for a completed/in-progress job. The CSV text is embedded directly in the result only when the job has at most 500 rows of results. Above that, to avoid burning your context window on a large export, the response instead carries `download_url` (a signed link valid for a few minutes, no API key needed), `expires_in_seconds` and `row_count` -- open the link yourself, fetch it with curl or your own HTTP client, or page through `get_results` (JSON) instead. Requires the `export` permission.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "get_results_csvArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      }
    }
  }

  arguments 13 lines copy
• estimate_enrich_list auth-required 12h ago

  Free price check for a person list. Never spends credits. Requires `contacts`. A person list looks up the email, phone and LinkedIn of people you name. Give `rows` (kind `name_domain`: objects with first_name, last_name, domain; kind `email`: objects with email) or `csv_text`, never both. `escalate_apollo` (name_domain only) also tries Apollo when Seamless finds nobody. Returns valid_count, invalid_count, duplicates, `invalid` (examples with reasons), prices, `already_found` (people your organization already found recently, skipped by start_enrich_list's `skip_known`), `min_credits_to_start` (the wallet must hold the price of one row), `worst_case_credits`, `balance`, `wallet_sufficient` and `next_step`.

  mcp-tool

  1234567891011121314151617181920212223242526272829303132333435363738394041424344

  {
    "type": "object",
    "title": "estimate_enrich_listArguments",
    "properties": {
      "kind": {
        "type": "string",
        "title": "Kind",
        "default": "name_domain"
      },
      "rows": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "object",
              "additionalProperties": true
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Rows",
        "default": null
      },
      "csv_text": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Csv Text",
        "default": null
      },
      "escalate_apollo": {
        "type": "boolean",
        "title": "Escalate Apollo",
        "default": false
      }
    }
  }

  arguments 44 lines copy
• get_enrich_list_csv auth-required 12h ago

  Return a person list's results as CSV. Requires `contacts`. One line per input row: status (found, not_found, skipped, error), reason, the found email, phone, LinkedIn URL, title and company, source and credits_charged. The CSV text is embedded when the list has at most 500 rows; above that the result carries `download_url` (a signed link valid for a few minutes, no API key needed) and `expires_in_seconds` instead. Works on a running list too: unfinished rows read queued or running.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "get_enrich_list_csvArguments",
    "required": [
      "list_id"
    ],
    "properties": {
      "list_id": {
        "type": "integer",
        "title": "List Id"
      }
    }
  }

  arguments 13 lines copy
• resume_job changes data auth-required never probed

  Resume a paused job and restart its senders. Requires `validate` permission. Cancelled jobs cannot be resumed. A 'done' job is a no-op: it reports `resumed: false, status: "done"` rather than a false success, and no workers are spawned for it.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "resume_jobArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      }
    }
  }

  arguments 13 lines copy
• list_jobs reads auth-required 18h ago

  List your email-validation jobs, most recent first. `limit` is the max jobs to return (default 20, clamped into 1..100). `offset` skips that many jobs for pagination (default 0, floored at 0 if negative) -- same contract as `get_results`. The response's `limit`/`offset` are always the EFFECTIVE values actually used, so a request outside range is never silently short without a signal: check the returned `limit`. Returns {jobs, limit, offset, has_more}; `has_more` is true when a next page actually has rows (fetch it with offset += limit), not merely whether this page happened to be full. No permission required.

  mcp-tool

  12345678910111213141516

  {
    "type": "object",
    "title": "list_jobsArguments",
    "properties": {
      "limit": {
        "type": "integer",
        "title": "Limit",
        "default": 20
      },
      "offset": {
        "type": "integer",
        "title": "Offset",
        "default": 0
      }
    }
  }

  arguments 16 lines copy
• cancel_job changes data auth-required never probed

  Cancel a job and refund credits for unsent leads. Requires `validate` permission. Marks the job cancelled so its workers stop, cancels any still-queued leads, and refunds one credit per charged lead not sent yet. Leads already sent still get a verdict after the bounce check; no-MX and send-failure refunds follow then. A job already 'done' or already 'cancelled' is a no-op: nothing changes and nothing is refunded again. Returns {job_id, cancelled, status, refunded, refunded_to, ...}. `refunded` is the wallet credit this call returned, not money moving to you; `refunded_to` names that wallet (always "wallet"). For a job funded through ZeroClick, `zeroclick_refund_note` explains the refund landed in Tuski's shared wallet and the buyer needs a separate ZeroClick-side refund.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "cancel_jobArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      }
    }
  }

  arguments 13 lines copy
• pause_job changes data auth-required never probed

  Pause a job so its senders stop claiming new leads. Requires `validate` permission. A job already 'done' or already 'cancelled' is a no-op: pausing never rewrites a terminal job's status (in particular, it never erases a cancelled job's terminal state). Returns {job_id, paused, status}. Call resume_job to continue sending.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "pause_jobArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      }
    }
  }

  arguments 13 lines copy
• start_enrich_list auth-required never probed

  Start a person list. Spends credits as people are found. Requires `contacts`. Same input as estimate_enrich_list, plus `skip_known` (people your organization already found recently are copied in free, not looked up) and a `name`. Charged per person found; a miss is free. The wallet must hold min_credits_to_start; the list stops when credits run out, keeping what it found. Invalid and repeated rows are dropped and counted. 5 per minute per key. Idempotency: the same `idempotency_key` returns the same list without starting it again; a different list under it is a conflict. Without one, a key is derived from every argument, so an identical call within 10 minutes is recognized. Keys starting with "auto:" are rejected. Runs in the background: call get_enrich_list until `finished`.

  mcp-tool

  12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273

  {
    "type": "object",
    "title": "start_enrich_listArguments",
    "properties": {
      "kind": {
        "type": "string",
        "title": "Kind",
        "default": "name_domain"
      },
      "name": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Name",
        "default": null
      },
      "rows": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "object",
              "additionalProperties": true
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Rows",
        "default": null
      },
      "csv_text": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Csv Text",
        "default": null
      },
      "skip_known": {
        "type": "boolean",
        "title": "Skip Known",
        "default": false
      },
      "escalate_apollo": {
        "type": "boolean",
        "title": "Escalate Apollo",
        "default": false
      },
      "idempotency_key": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Idempotency Key",
        "default": null
      }
    }
  }

  arguments 73 lines copy
• cancel_enrich_list auth-required never probed

  Stop a running person list. Requires `contacts`. Rows not yet looked up are skipped and never charged; a person already found stays found and charged. A list that already finished is left as it is, so cancelling twice is harmless. Returns the list's status and counts.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "cancel_enrich_listArguments",
    "required": [
      "list_id"
    ],
    "properties": {
      "list_id": {
        "type": "integer",
        "title": "List Id"
      }
    }
  }

  arguments 13 lines copy
• estimate_company_search reads auth-required 18h ago

  Free price check for an AI Web Search company search. Never reserves credits. `effort` is one of "scout" (cheapest, fastest, fewest results), "standard" (default), or "deep" (highest quality, may be unavailable on this instance). `target` is the number of verified companies you want (defaults to the effort tier's own default, capped at its `max_target`). Add `icp_text` (and `rules_json`) to also get its `filters` and `warnings`. Returns `tier`, `target`, `max_target`, `base_fee`, `result_price`, `reserved_credits` (the exact amount `start_company_search` would reserve: `base_fee` + `target` x `result_price`; a search charges `base_fee` + `result_price` per verified company it delivers and refunds the rest), `wallet_balance`, `wallet_sufficient`, and a `next_step`. Requires the `discovery` permission.

  mcp-tool

  12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273

  {
    "type": "object",
    "$defs": {
      "EffortTier": {
        "enum": [
          "scout",
          "standard",
          "deep"
        ],
        "type": "string",
        "title": "EffortTier"
      }
    },
    "title": "estimate_company_searchArguments",
    "properties": {
      "effort": {
        "$ref": "#/$defs/EffortTier",
        "default": "standard"
      },
      "target": {
        "anyOf": [
          {
            "type": "integer"
          },
          {
            "type": "null"
          }
        ],
        "title": "Target",
        "default": null
      },
      "icp_text": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Icp Text",
        "default": null
      },
      "provider": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Provider",
        "default": null
      },
      "rules_json": {
        "anyOf": [
          {
            "type": "object",
            "additionalProperties": true
          },
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Rules Json",
        "default": null
      }
    }
  }

  arguments 73 lines copy
• start_company_search moves money auth-required never probed

  Start an AI Web Search company search. Reserves credits up front. `icp_text` is the ideal customer profile in plain language; `effort` and `target` set price and quality (estimate_company_search quotes them). Filters are verified on each company's site; only passing ones are charged. Companies your organization already received are skipped unless `rules_json` sets "include_previous": true. `example_companies` (up to 3) refine matching. Needs `discovery`; 10 per minute per key. Idempotency: the same `idempotency_key` returns the existing search without reserving again; a different payload under it is a conflict. Without one, a key is derived from every argument, so a repeat within 10 minutes is recognized. Keys starting with "auto:" are rejected. Returns the status plus `created` and `idempotency`; then call wait_for_company_search.

  mcp-tool

  123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293

  {
    "type": "object",
    "$defs": {
      "EffortTier": {
        "enum": [
          "scout",
          "standard",
          "deep"
        ],
        "type": "string",
        "title": "EffortTier"
      }
    },
    "title": "start_company_searchArguments",
    "required": [
      "icp_text"
    ],
    "properties": {
      "effort": {
        "$ref": "#/$defs/EffortTier",
        "default": "standard"
      },
      "target": {
        "anyOf": [
          {
            "type": "integer"
          },
          {
            "type": "null"
          }
        ],
        "title": "Target",
        "default": null
      },
      "icp_text": {
        "type": "string",
        "title": "Icp Text"
      },
      "provider": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Provider",
        "default": null
      },
      "rules_json": {
        "anyOf": [
          {
            "type": "object",
            "additionalProperties": true
          },
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Rules Json",
        "default": null
      },
      "idempotency_key": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Idempotency Key",
        "default": null
      },
      "example_companies": {
        "anyOf": [
          {
            "type": "array",
            "items": {}
          },
          {
            "type": "null"
          }
        ],
        "title": "Example Companies",
        "default": null
      }
    }
  }

  arguments 93 lines copy
• wait_for_company_search reads auth-required 7h ago

  Wait on the server for a company search to finish, instead of polling yourself. Checks the search's status every few seconds until it is terminal (succeeded, failed, or canceled), or until `max_wait_seconds` elapses, whichever comes first. `max_wait_seconds` (default 25) is clamped into 1..50: this call itself never blocks longer than that, and a still-running search simply comes back with `done: false` rather than an error. Returns everything `get_company_search` returns, plus `poll_after_seconds` and `next_step`. Requires the `discovery` permission.

  mcp-tool

  123456789101112131415161718

  {
    "type": "object",
    "title": "wait_for_company_searchArguments",
    "required": [
      "search_id"
    ],
    "properties": {
      "search_id": {
        "type": "integer",
        "title": "Search Id"
      },
      "max_wait_seconds": {
        "type": "integer",
        "title": "Max Wait Seconds",
        "default": 25
      }
    }
  }

  arguments 18 lines copy
• export_companies reads auth-required never probed

  Export a company search's delivered companies (the rows get_companies lists) as CSV. The last columns are `billable`, `verification_status` and `unverified_filters`. The CSV text is embedded directly in the result only when the search has at most 500 result rows. Above that, to avoid burning your context window on a large export, the response instead carries `download_url` (a signed link valid for a few minutes, no API key needed), `expires_in_seconds` and `row_count` -- open the link yourself, fetch it with curl or your own HTTP client, or page through `get_companies` (JSON) instead. Requires the `discovery` permission.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "export_companiesArguments",
    "required": [
      "search_id"
    ],
    "properties": {
      "search_id": {
        "type": "integer",
        "title": "Search Id"
      }
    }
  }

  arguments 13 lines copy
• cancel_company_search changes data auth-required never probed

  Cancel a running company search. Settles the reservation against results delivered so far. Refunds whatever part of the reservation was not earned by verified companies already delivered. `cancelled` is false when the search was already terminal (succeeded/failed/already canceled) -- a second cancel is a harmless no-op, not an error, and nothing is released again. `released_credits` is the reservation's total refund to date (0 if nothing was ever refunded). Requires the `discovery` permission.

  mcp-tool

  12345678910111213

  {
    "type": "object",
    "title": "cancel_company_searchArguments",
    "required": [
      "search_id"
    ],
    "properties": {
      "search_id": {
        "type": "integer",
        "title": "Search Id"
      }
    }
  }

  arguments 13 lines copy
• find_contacts changes data auth-required never probed

  Start a free people search for the companies of a finished company search. `search_id` is the id from `start_company_search`/`get_company_search`. `candidate_ids` selects which of that search's companies to search (from `get_companies`). Optional `titles`/`seniorities`/`exclude_titles` narrow who counts as a match. Always uses the apollo provider. Nothing is charged here; only `unlock_contacts` spends credit. Returns `job_id` (pass it to `get_contacts`, `estimate_contact_unlock` and `unlock_contacts`) and `status`. Idempotency: pass `idempotency_key` for safe retries; a repeat with the same key reuses the same job. Send none and one is derived from your arguments; a repeat within 10 minutes is recognized the same way. Requires the `contacts` permission.

  mcp-tool

  123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778

  {
    "type": "object",
    "title": "find_contactsArguments",
    "required": [
      "search_id",
      "candidate_ids"
    ],
    "properties": {
      "titles": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "string"
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Titles",
        "default": null
      },
      "search_id": {
        "type": "integer",
        "title": "Search Id"
      },
      "seniorities": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "string"
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Seniorities",
        "default": null
      },
      "candidate_ids": {
        "type": "array",
        "items": {
          "type": "integer"
        },
        "title": "Candidate Ids"
      },
      "exclude_titles": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "string"
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Exclude Titles",
        "default": null
      },
      "idempotency_key": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Idempotency Key",
        "default": null
      }
    }
  }

  arguments 78 lines copy
• get_companies reads auth-required 18h ago

  List one page of a company search's delivered companies, paginated. A search lists the companies it verified, the ones it charges for, and free `size_unverified` ones (headcount not stated; see `verification`). A search from an earlier version (`verified_results` false) lists every candidate it found. `limit` (default 25, clamped into 1..100) and `offset` (default 0, floored at 0) page through them in the order found. Rows describe companies: `email`, `phone` and `person_name` are always null; use `find_contacts` for people, then `validate_emails` on their emails. `outcome_reason` is a short code such as `accepted`. Returns `{search_id, rows, total, limit, offset, has_more, next_offset}`; `next_offset` is the offset to request next, or null once there is nothing left. Requires the `discovery` permission.

  mcp-tool

  1234567891011121314151617181920212223

  {
    "type": "object",
    "title": "get_companiesArguments",
    "required": [
      "search_id"
    ],
    "properties": {
      "limit": {
        "type": "integer",
        "title": "Limit",
        "default": 25
      },
      "offset": {
        "type": "integer",
        "title": "Offset",
        "default": 0
      },
      "search_id": {
        "type": "integer",
        "title": "Search Id"
      }
    }
  }

  arguments 23 lines copy
• validate_emails sends messages auth-required never probed

  Start an email-validation job. Spends credits and sends real probe emails. Provide exactly one of `csv_text` or `emails`; supplying both is rejected. Requires `validate` and deducts 1 credit per email accepted for testing (passes syntax and domain checks). No-MX addresses and our own send failures are refunded when the job closes; a real bounce is a paid result. Tuski writes the probe text itself. Call `get_credits` first. Pass `idempotency_key` for safe retries: the same key returns the existing job instead of charging twice; a different list under the same key is a conflict. Send none and one is derived from the account, org and list; a repeat within 10 minutes is recognized the same way. Keys starting with "auto:" are rejected. Call wait_for_job next. fresh_probe skips remembered results.

  mcp-tool

  1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950

  {
    "type": "object",
    "title": "validate_emailsArguments",
    "properties": {
      "emails": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "string"
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Emails",
        "default": null
      },
      "csv_text": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Csv Text",
        "default": null
      },
      "fresh_probe": {
        "type": "boolean",
        "title": "Fresh Probe",
        "default": false
      },
      "idempotency_key": {
        "anyOf": [
          {
            "type": "string"
          },
          {
            "type": "null"
          }
        ],
        "title": "Idempotency Key",
        "default": null
      }
    }
  }

  arguments 50 lines copy
• wait_for_contacts reads auth-required 4h ago

  Wait on the server for a set of unlock requests to finish. `request_ids` are ids `unlock_contacts` returned. Checks every few seconds until every request reaches a terminal status ("completed", "no_result" or "unknown"), or until `max_wait_seconds` (default 25, clamped into 1..50) elapses, whichever comes first. Returns `requests` (each with its own `status`, and for a `verify=true` unlock the `verification_job_id` of its email check once created), `done`, and `poll_after_seconds`. If `done` is false, wait `poll_after_seconds` and call this again. Once done, call `get_contacts` to read the revealed fields. Requires the `contacts` permission.

  mcp-tool

  123456789101112131415161718192021

  {
    "type": "object",
    "title": "wait_for_contactsArguments",
    "required": [
      "request_ids"
    ],
    "properties": {
      "request_ids": {
        "type": "array",
        "items": {
          "type": "integer"
        },
        "title": "Request Ids"
      },
      "max_wait_seconds": {
        "type": "integer",
        "title": "Max Wait Seconds",
        "default": 25
      }
    }
  }

  arguments 21 lines copy
• wait_for_job reads auth-required 4h ago

  Wait on the server for a job to finish, instead of polling yourself. Checks the job's status every few seconds until it is 'done' or 'cancelled', or until `max_wait_seconds` (default 25, clamped into 1..50) elapses, whichever comes first. Returns everything get_job_status returns, plus `done`, `poll_after_seconds` (0 when done, a suggested next wait otherwise: longer, up to 15 minutes, for a job that will take hours) and `next_step`. If `done` is false, wait `poll_after_seconds` and call this again. Once done, call get_results. No permission required.

  mcp-tool

  123456789101112131415161718

  {
    "type": "object",
    "title": "wait_for_jobArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      },
      "max_wait_seconds": {
        "type": "integer",
        "title": "Max Wait Seconds",
        "default": 25
      }
    }
  }

  arguments 18 lines copy
• estimate_contact_unlock reads auth-required 4h ago

  Free price check for unlocking contact fields. Never spends credit. `job_id` is the id `find_contacts` returned. Optional `selections` (a list of `{person_id, fields}`, fields "email"/"linkedin"/"phone") adds `estimated_cost_upper_bound`, `wallet_sufficient` and `verify_cost_upper_bound` for that choice. Returns `costs` (per field), `email_unverified_price` (a guessed or personal email), `verify_price_per_email` (per email checked by `unlock_contacts` with `verify=true`), `phone_enabled`, `balance` and `unlockable_people`. Owned fields and LinkedIn URLs the search returned are free on the real call, so it can charge less, never more. Paid from the organization's credit balance. Requires `contacts`.

  mcp-tool

  1234567891011121314151617181920212223242526272829

  {
    "type": "object",
    "title": "estimate_contact_unlockArguments",
    "required": [
      "job_id"
    ],
    "properties": {
      "job_id": {
        "type": "integer",
        "title": "Job Id"
      },
      "selections": {
        "anyOf": [
          {
            "type": "array",
            "items": {
              "type": "object",
              "additionalProperties": true
            }
          },
          {
            "type": "null"
          }
        ],
        "title": "Selections",
        "default": null
      }
    }
  }

  arguments 29 lines copy
• get_credits reads auth-required never probed

  Check the credit balance for this API key's organization. Returns credits (also payg_credits and total), the organization wallet balance -- the SAME wallet validate_emails spends from (a Team-org key reports that Team's wallet, not the caller's personal one). monthly_tokens and free_credits_remaining are always 0 now that every grant lands in the wallet, kept for compatibility. Also returns organization_id, organization_kind, wallet, plan_type and subscription_status. validate_emails deducts 1 credit per email accepted for testing (syntax and domain checks pass); no-MX addresses and our own send failures are refunded at close, but a real bounce is a paid result. Call this first to avoid failing mid-run. Any valid API key works (no special permission required).

  mcp-tool

  12345

  {
    "type": "object",
    "title": "get_creditsArguments",
    "properties": {}
  }

  arguments 5 lines copy

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

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.

Served from the same domain, which is what was measured. Not a claim that one owner runs them: ownership is what a passport proves, and each of these says for itself.

• actionlayer.pay.zeroclick.io actionlayer.pay.zeroclick.io mcp · 0 tools