_ registry / mcp streamable-http

korus-marketplace-mcp

https://www.korusmusic.co

Registry code: 902378de792d774e

api record
endpoint
https://www.korusmusic.co/api/mcp
protocol
streamable-http ·2025-06-18
authentication
none observed
public key
none — nobody has proven they own this listing
karma
0 · newcomer
reachable
unknown
uptime
—
latency
—

last good check

priced tools
0

of 19 tools

_ used through this hub 30 days

The one measurement on this page that an operator cannot produce by editing a file on its own server: somebody else chose it, and paid to. Read the accounts before the calls — volume from one account is one relationship, and calling yourself is the cheap half. Both are what the ranking is built from, printed so the order can be checked rather than taken on trust.

accounts
0

distinct, expensive to fake

calls served
0

successful, last 30 days

_ what it can do 19 tools
19 never probed 0 of 19 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_licensed_asset unknown never probed

    Resolve presigned full-quality download URL(s) for every asset a license covers. Authorized as the license's grantee (session, session+linked wallet, or a signature-verified wallet) — a non-grantee, revoked, or unknown license id all 404 identically. (v1 API: GET /api/v1/licenses/:licenseId/assets) Requires an authenticated session — run session_challenge then session_authenticate first. This is how you TAKE DELIVERY of a stem after a purchase: call it with the licenseId returned by license_stem (session rail) or license_settle (x402 rail) to resolve a presigned, fetchable URL for every asset that license covers — links only, this tool never returns audio bytes. Each url expires in its expiresInSec seconds; call this tool again for a fresh one rather than caching a stale link. A 404 means this license id is not yours (or has been revoked) — do not retry, the answer will not change.

    mcp-tool

    {
  "type": "object",
  "required": [
    "licenseId"
  ],
  "properties": {
    "licenseId": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • search_stems unknown never probed

    Search/browse active listings. (v1 API: GET /api/v1/catalog) Prices: quote `priceDisplay` verbatim; `priceMinor` is in minor units (usd cents, usdc 6-decimal micro-units) — never divide a usdc `priceMinor` by 100.

    mcp-tool

    {
  "type": "object",
  "required": [
    "sort",
    "limit"
  ],
  "properties": {
    "q": {
      "type": "string"
    },
    "key": {
      "type": "string"
    },
    "kind": {
      "enum": [
        "vocals",
        "drums",
        "bass",
        "synths",
        "guitar",
        "other"
      ],
      "type": "string"
    },
    "sort": {
      "enum": [
        "relevance",
        "newest",
        "price_asc",
        "price_desc",
        "popularity"
      ],
      "type": "string"
    },
    "limit": {
      "type": "number"
    },
    "artist": {
      "type": "string"
    },
    "bpmMax": {
      "type": "number"
    },
    "bpmMin": {
      "type": "number"
    },
    "cursor": {
      "type": "string"
    },
    "priceMax": {
      "type": "number"
    },
    "priceMin": {
      "type": "number"
    },
    "subjectType": {
      "enum": [
        "stem",
        "stem_pack"
      ],
      "type": "string"
    },
    "durationMaxSec": {
      "type": "number"
    },
    "durationMinSec": {
      "type": "number"
    }
  }
}
    arguments 70 lines
  • get_listing unknown never probed

    Fetch one listing's public detail. (v1 API: GET /api/v1/listings/:listingId) Prices: quote `priceDisplay` verbatim; `priceMinor` is in minor units (usd cents, usdc 6-decimal micro-units) — never divide a usdc `priceMinor` by 100.

    mcp-tool

    {
  "type": "object",
  "required": [
    "listingId"
  ],
  "properties": {
    "listingId": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • preview_stem unknown never probed

    Resolve presigned preview URL(s) for a listing (previewStorageKey only, never full-quality bytes). (v1 API: GET /api/v1/previews/:listingId)

    mcp-tool

    {
  "type": "object",
  "required": [
    "listingId"
  ],
  "properties": {
    "listingId": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • license_stem unknown never probed

    Purchase a license for a listing. Session-rail: idempotent, creates a pending Order; the License is granted asynchronously once fulfill_order processes payment. Anonymous (no session): the SAME path answers a stateless x402 402 challenge (no X-PAYMENT header) or verifies-and-settles a payment (X-PAYMENT header present) — see responses. (v1 API: POST /api/v1/licenses/:listingId) Requires an authenticated session — run session_challenge then session_authenticate first. Buys as the signed-in user via the session rails: always works for free listings (priceMinor 0); card checkout for priced listings depends on a deployment flag (STRIPE_CHECKOUT_ENABLED) that is currently disabled in production — a priced usd purchase then returns a structured 503 with guidance instead of a bare error. Priced listings are paid with USDC via license_quote (get the 402 challenge) then license_settle (submit the wallet-signed payment).

    mcp-tool

    {
  "type": "object",
  "required": [
    "listingId",
    "idempotencyKey"
  ],
  "properties": {
    "listingId": {
      "type": "string"
    },
    "idempotencyKey": {
      "type": "string"
    }
  }
}
    arguments 15 lines
  • verify_rights unknown never probed

    Check whether an asset's provenance traces to a License (Track.licenseId) or an original upload (public, no session). (v1 API: GET /api/v1/rights/:assetId)

    mcp-tool

    {
  "type": "object",
  "required": [
    "assetId"
  ],
  "properties": {
    "assetId": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • verify_license unknown never probed

    Verify a license key is currently active (public, no session). (v1 API: GET /api/v1/licenses/verify/:licenseKey)

    mcp-tool

    {
  "type": "object",
  "required": [
    "licenseKey"
  ],
  "properties": {
    "licenseKey": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • get_job unknown never probed

    Poll a job you created (e.g. from remixCreate). Once succeeded, result.tracks lists each produced track with a short-lived download URL — never storageKey — plus a studioUrl to hear them together (no combined mix exists). Also carries result.files[] (suggested filename + mimeType per track, same URLs) and result.previewUrl (a tokenized, no-login player page, 15 min TTL) — desktop/web clients should open previewUrl or studioUrl; terminal clients should download files[]. A foreign or unknown job id both 404 identically. (v1 API: GET /api/v1/jobs/:jobId) Requires an authenticated session — run session_challenge then session_authenticate first. Poll a remix job; status is queued|running|succeeded|failed; on success result.tracks[] are presigned download URLs (expire in expiresInSec) plus a studio deep link; 404 means not your job — do not retry. On success: Desktop/web users -> open result.previewUrl (plays in the browser, no login, 15 min) or result.studioUrl (full editor). Terminal clients (e.g. Claude Code) -> download each result.files[].url into the directory the user names, using files[].name; URLs expire in expiresInSec.

    mcp-tool

    {
  "type": "object",
  "required": [
    "jobId"
  ],
  "properties": {
    "jobId": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • list_my_library unknown never probed

    List your reusable audio library: every stem you uploaded/generated/split, plus every stem you hold an unrevoked marketplace license for (including one granted anonymously to your linked wallet). Use assetId from a row as remix_stem's source.assetId. Session-authed (D7 — the wallet you signed in with IS this account); a signed-in caller with nothing yet gets items: [], never a 404. (v1 API: GET /api/v1/me/library) Requires an authenticated session — run session_challenge then session_authenticate first. "Your" library = every stem the wallet you signed in with uploaded/generated, PLUS every stem it holds an unrevoked marketplace license for. Use a row's assetId as remix_stem/create_track's sources[].assetId (or the legacy remix_stem source.assetId).

    mcp-tool

    {
  "type": "object",
  "properties": {
    "limit": {
      "type": "number"
    },
    "cursor": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • list_my_licenses unknown never probed

    List the licenses you hold (receipts), newest first — the same union GET /api/licenses reads (your session grants plus any granted to your linked wallet). Use licenseId from a row as remix_stem's source.licenseId, or get_licensed_asset's licenseId. Quote purchase.priceDisplay, never divide a usdc purchase.amountMinor by 100. Session-authed. (v1 API: GET /api/v1/me/licenses) Requires an authenticated session — run session_challenge then session_authenticate first. "Your" licenses = every receipt granted to the wallet you signed in with (including one granted anonymously before you linked/logged in with it). Use a row's id as remix_stem/create_track's sources[].licenseId (or the legacy remix_stem source.licenseId). Quote purchase.priceDisplay verbatim — never divide a usdc purchase.amountMinor by 100.

    mcp-tool

    {
  "type": "object",
  "properties": {
    "limit": {
      "type": "number"
    },
    "cursor": {
      "type": "string"
    }
  }
}
    arguments 11 lines
  • list_my_projects unknown never probed

    List the projects you own, most recently updated first, each with a studioUrl deep link and its track count. Session-authed. (v1 API: GET /api/v1/me/projects) Requires an authenticated session — run session_challenge then session_authenticate first. "Your" projects = every project the wallet you signed in with owns. Each row's studioUrl opens it.

    mcp-tool

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

    Get the x402 payment challenge to license a listing with a crypto wallet (USDC on Base). THE path for a priced listing: card checkout (license_stem) depends on a deployment flag that is off in production, so license_quote -> a wallet agent pays -> license_settle is how priced listings get paid. Returns HTTP 402 plus the x402 PaymentRequirements (amount, payTo treasury, asset, network, EIP-712 domain). The quoted maxAmountRequired includes a sub-cent listing-binding salt (e.g. 0.5003 USDC for a $0.50 listing) — pay EXACTLY that amount; a proof signed for a different amount or listing is rejected before settlement and re-challenged. Hand this challenge to a wallet agent (e.g. Base MCP) to sign an EIP-3009 authorization, then call license_settle with the resulting X-PAYMENT proof. No KORUS account/session needed.

    mcp-tool

    {
  "type": "object",
  "required": [
    "listingId"
  ],
  "properties": {
    "listingId": {
      "type": "string",
      "description": "The listing to price a license for."
    }
  },
  "additionalProperties": false
}
    arguments 13 lines
  • license_settle unknown never probed

    Settle a wallet-paid license by submitting a signed x402 proof (the base64 X-PAYMENT header a wallet agent like Base MCP produced from license_quote's challenge). THE second half of the license_quote path for priced listings. On success returns the Order + granted License (license.id/licenseKey) tied to the paying wallet — no KORUS session involved. A rejected/expired proof returns 402; the License is granted via the same fulfill_order job every rail uses.

    mcp-tool

    {
  "type": "object",
  "required": [
    "listingId",
    "xPayment"
  ],
  "properties": {
    "xPayment": {
      "type": "string",
      "description": "The base64-encoded X-PAYMENT proof (signed EIP-3009 authorization) from the wallet agent."
    },
    "listingId": {
      "type": "string",
      "description": "The same listing license_quote was called for."
    }
  },
  "additionalProperties": false
}
    arguments 18 lines
  • mint_prepare unknown never probed

    Prepare an on-chain mint of a licensed remix on KOR Protocol. Validates the caller owns a rights-clean remix and returns an UNSIGNED transaction { to, data, value, chainId } for a wallet agent (e.g. Base MCP) to broadcast. After broadcasting, call mint_confirm with the txHash. Requires an authenticated session (the remix owner) — run session_challenge then session_authenticate first.

    mcp-tool

    {
  "type": "object",
  "required": [
    "trackId",
    "recipientAddress"
  ],
  "properties": {
    "trackId": {
      "type": "string",
      "description": "The remix Track to mint."
    },
    "metadataURI": {
      "type": "string",
      "description": "Optional ipfs:// token metadata URI."
    },
    "recipientAddress": {
      "type": "string",
      "description": "Wallet that will own the minted NFT (0x…)."
    }
  },
  "additionalProperties": false
}
    arguments 22 lines
  • mint_confirm unknown never probed

    Confirm a broadcast mint: given the txHash the wallet agent returned, KORUS reads the on-chain receipt, decodes the minted tokenId, and records an immutable MintRecord (idempotent on txHash). Returns { tokenId, contractAddress, txHash, explorer }. Requires the server session (the remix owner).

    mcp-tool

    {
  "type": "object",
  "required": [
    "trackId",
    "txHash"
  ],
  "properties": {
    "txHash": {
      "type": "string",
      "description": "The broadcast transaction hash (0x…)."
    },
    "trackId": {
      "type": "string",
      "description": "The Track that was minted (same as mint_prepare)."
    }
  },
  "additionalProperties": false
}
    arguments 18 lines
  • session_challenge unknown never probed

    Start a KORUS login for a wallet you control. Returns `challenge.message` — hand it to a wallet agent's sign tool (e.g. Base MCP `sign` with type "personal_sign" and data {message}), which returns an approvalUrl for the user to approve in their Base Account; poll get_request_status(requestId) until it completes to get the signature. Then call session_authenticate with the SAME address and timestamp plus that signature. The challenge expires in 10 minutes. The wallet must already be linked to a KORUS account (log in with it via the Para modal once, or link it in Settings) — KORUS never creates an account from a bare signature.

    mcp-tool

    {
  "type": "object",
  "required": [
    "address"
  ],
  "properties": {
    "address": {
      "type": "string",
      "description": "The 0x wallet address to log in as."
    }
  },
  "additionalProperties": false
}
    arguments 13 lines
  • session_authenticate unknown never probed

    Complete the login started by session_challenge: submit the wallet's personal_sign signature over the challenge message. On success the session is held in memory for this server's lifetime and the authed tools (license_stem, get_licensed_asset, remix_stem, create_track, get_job, mint_prepare, mint_confirm, list_my_library, list_my_licenses, list_my_projects) start working. The session token itself is never returned. Fails if the challenge expired, the signature doesn't match the address, or the address isn't linked to a KORUS account.

    mcp-tool

    {
  "type": "object",
  "required": [
    "address",
    "timestamp",
    "signature"
  ],
  "properties": {
    "address": {
      "type": "string",
      "description": "The SAME address passed to session_challenge."
    },
    "signature": {
      "type": "string",
      "description": "The 0x signature the wallet agent produced over challenge.message."
    },
    "timestamp": {
      "type": "string",
      "description": "The SAME challenge.timestamp returned by session_challenge."
    }
  },
  "additionalProperties": false
}
    arguments 23 lines
  • remix_stem unknown never probed

    Generate ONE new stem (targetKind) that follows an existing stem's rhythm/tempo/key, and attach it next to the untouched source — stem IN, stem OUT, the source itself is never modified. Works best when the source is drums, vocals, or otherwise rhythmic/transient; a sustained pad or synth pad conditions weakly (the model mostly re-synthesises it) — say so if a caller asks for that. `instruction` steers WHAT is added (e.g. "add a driving bassline"); omit it for the model's own choice from `prompt` alone. WAV sources only today — a non-WAV source (e.g. an mp3 upload) fails the job with a clear error rather than silently falling back to an unconditioned generation. Pass ONE anchor source via `sources: [{assetId|licenseId|trackId}]` (1-4 entries accepted; only the first — the anchor — actually conditions the output today) or the legacy singular `source`/`sourceTrackId`; omit projectId with `sources`/`source` to auto-create a fresh project. Returns a job to poll with get_job; on success the new stem lands in the project's studio (studioUrl) next to the source — never a combined mix. Requires an authenticated session — run session_challenge then session_authenticate first.

    mcp-tool

    {
  "type": "object",
  "required": [
    "prompt",
    "targetKind"
  ],
  "properties": {
    "params": {
      "type": "object",
      "properties": {
        "bars": {
          "type": "number"
        },
        "name": {
          "type": "string"
        }
      },
      "additionalProperties": false
    },
    "prompt": {
      "type": "string",
      "description": "Style/genre prompt for the new stem."
    },
    "source": {
      "type": "object",
      "properties": {
        "assetId": {
          "type": "string"
        },
        "trackId": {
          "type": "string"
        },
        "licenseId": {
          "type": "string"
        }
      },
      "description": "Legacy singular alias for sources[0]. Mutually exclusive with sources/sourceTrackId.",
      "additionalProperties": false
    },
    "sources": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "assetId": {
            "type": "string",
            "description": "An asset you own or hold a license for — from list_my_library."
          },
          "trackId": {
            "type": "string",
            "description": "A track already attached to projectId."
          },
          "licenseId": {
            "type": "string",
            "description": "A license you hold — from list_my_licenses."
          }
        },
        "additionalProperties": false
      },
      "maxItems": 4,
      "minItems": 1,
      "description": "1-4 sources, exactly one key each; the FIRST is the conditioning anchor (the rest attach for arrangement only — see create_track for a tool that actually blends multiple sources into one track)."
    },
    "projectId": {
      "type": "string",
      "description": "Omit to auto-create a fresh project for a sources/source-based call; required when sourceTrackId (or a trackId source) is used."
    },
    "targetKind": {
      "enum": [
        "vocals",
        "drums",
        "bass",
        "synths",
        "guitar",
        "other"
      ],
      "type": "string",
      "description": "The kind of stem to generate — must differ from the source's own kind (this ships as a stem *addition*, not an \"improve this same stem\" restyle). targetKind \"vocals\" is not yet supported with sources (the lyric model overrides the source audio; measured 2026-09-12) — the server rejects it."
    },
    "instruction": {
      "type": "string",
      "description": "≤200 chars. Steers what gets added (e.g. instrument/part named). Optional."
    },
    "sourceTrackId": {
      "type": "string",
      "description": "Legacy MP-08 field — a track already attached to projectId. Mutually exclusive with source/sources."
    }
  },
  "additionalProperties": false
}
    arguments 90 lines
  • create_track unknown never probed

    Generate ONE mixdown track from one or more existing stems — stems IN, ONE finished track OUT (the opposite shape of remix_stem, which only ever adds a single new stem). The result CONTAINS the source(s) plus whatever was added (`task: "complete"`, the default) or a restyled version of the source (`task: "cover"`) — either way, treat the returned track as the new mix and mute/remove the original source tracks in the studio rather than layering both. Conditioning is strongest from a SINGLE rhythmic or vocal source; passing several sources as one mix (`sources` accepts 1-4) conditions only loosely — the model follows the overall mix weakly, it does not cleanly preserve each part. Output length always equals the (first/anchor) source's length. WAV sources only today — a non-WAV source fails the job with a clear error rather than silently falling back to plain text2music. `instruction` (complete only, ≤200 chars) steers what's added — SR-10: with `task:"complete"` and no `instruction`, one is DERIVED from the anchor's own kind (e.g. a bass anchor completes with drums/synth/keyboard/percussion, never the anchor's own part and never vocals by default — the LM is off for conditioned takes); pass your own `instruction` to steer what gets added instead, including asking for vocals. `strength` (cover only, 0-1) sets how far the restyle departs from the source. Returns a job to poll with get_job; on success the track lands in the project's studio (studioUrl). Requires an authenticated session — run session_challenge then session_authenticate first.

    mcp-tool

    {
  "type": "object",
  "required": [
    "prompt",
    "sources"
  ],
  "properties": {
    "task": {
      "enum": [
        "complete",
        "cover"
      ],
      "type": "string",
      "description": "\"complete\" (default): adds new parts on top of the source(s). \"cover\": restyles the source(s) instead of adding to them."
    },
    "params": {
      "type": "object",
      "properties": {
        "bars": {
          "type": "number"
        },
        "name": {
          "type": "string"
        },
        "lyrics": {
          "type": "string"
        }
      },
      "additionalProperties": false
    },
    "prompt": {
      "type": "string",
      "description": "Style/genre prompt for the generated track."
    },
    "sources": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "assetId": {
            "type": "string",
            "description": "An asset you own or hold a license for — from list_my_library."
          },
          "trackId": {
            "type": "string",
            "description": "A track already attached to projectId."
          },
          "licenseId": {
            "type": "string",
            "description": "A license you hold — from list_my_licenses."
          }
        },
        "additionalProperties": false
      },
      "maxItems": 4,
      "minItems": 1,
      "description": "1-4 sources, exactly one key each. A SINGLE source (especially rhythmic/vocal) conditions best; more than one is a weak, loosely-followed mix."
    },
    "strength": {
      "type": "number",
      "description": "0-1. cover only — how far the restyle departs from the source."
    },
    "projectId": {
      "type": "string",
      "description": "Omit to auto-create a fresh project (only when every source resolves via assetId/licenseId — a trackId source requires projectId, same as remix_stem)."
    },
    "instruction": {
      "type": "string",
      "description": "≤200 chars. complete only — steers what gets added."
    }
  },
  "additionalProperties": false
}
    arguments 73 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.

_ for your README measured, not declared

measured by brick.blue

[![measured by brick.blue](https://brick.blue/api/v1/agents/902378de792d774e/badge.svg)](https://brick.blue/agent/902378de792d774e)

The picture says what this hub measured — the access class, how many tools it called and whether they answered — and refreshes hourly. Own the domain? Prove it and the listing carries a verified badge here too: passport.

_ 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.