View as Markdown

All MCP tools

recordings_list

List recordings

List standalone recordings (raw source media) — the drawer, not the editorial unit. Use for source/ingest/readiness asks ("latest searchable session", "what uploads are ready?") or to inspect and select recordings inside a project (project_id). For one Project reel whose useful footage spans separate recordings rather than one parent Episode, inspect selected recording pixels directly with recording_sample_frames when needed, then pass recording_id ranges directly to one episode_create without derived_from_id. Never create a throwaway Episode merely to inspect a Recording. `query` matches title, project, or who is in it ("the latest with Andrew" matches Andrew as a speaker). PAGINATED, newest first: one call returns up to 20 rows and says when more exist; continue with the returned next_cursor (same query/project_id) as `cursor`. One listing walk reaches at most the newest 50 recordings — anything older is reachable only by narrowing with query or project_id, and the result states which case you are in. Model rows carry both the exact tool-only recording_id and the episode-of-one episode_id, plus captured_at, searchability status/notes, and a compact indexed description. Prefer the newest `searchability_status:"searchable"` row for latest-session asks; episode_pending means the Recording may not have an Episode-of-one, not that direct source-time frame inspection must wait. Not for ordinary composition browsing (episodes_list) and not for searching what was SAID (transcript_search). Read-only.

Surface: Default model and app

Product release 1dce717408a55fac1047e7d779177fa3ef8ede51
Public contract commit 1ce47f7857ae3d9b6933fe83d5c33de3ae5efc64
Public contract SHA-256 241d47d25e4b4c1dae9d2263793874d06fc4ba7bc3c78d4cf2bc63fbb32ed7dc
Captured 2026-09-24T16:39:30.100Z. Review or improve this contract on GitHub.

These blocks show the complete MCP descriptors captured from serving Rails. Security schemes and resource URIs can differ by connected host.

Default model descriptor

{
  "name": "recordings_list",
  "title": "List recordings",
  "description": "List standalone recordings (raw source media) — the drawer, not the editorial unit. Use for source/ingest/readiness asks (\"latest searchable session\", \"what uploads are ready?\") or to inspect and select recordings inside a project (project_id). For one Project reel whose useful footage spans separate recordings rather than one parent Episode, inspect selected recording pixels directly with recording_sample_frames when needed, then pass recording_id ranges directly to one episode_create without derived_from_id. Never create a throwaway Episode merely to inspect a Recording. `query` matches title, project, or who is in it (\"the latest with Andrew\" matches Andrew as a speaker). PAGINATED, newest first: one call returns up to 20 rows and says when more exist; continue with the returned next_cursor (same query/project_id) as `cursor`. One listing walk reaches at most the newest 50 recordings — anything older is reachable only by narrowing with query or project_id, and the result states which case you are in. Model rows carry both the exact tool-only recording_id and the episode-of-one episode_id, plus captured_at, searchability status/notes, and a compact indexed description. Prefer the newest `searchability_status:\"searchable\"` row for latest-session asks; episode_pending means the Recording may not have an Episode-of-one, not that direct source-time frame inspection must wait. Not for ordinary composition browsing (episodes_list) and not for searching what was SAID (transcript_search). Read-only.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "query": {
        "type": "string",
        "description": "Optional case-insensitive filter — returns only recordings whose TITLE, a composited SOURCE TITLE, or a diarized SPEAKER NAME contains this text (substring). The filter scans the FULL catalog before the newest-50 listing window applies, so it also reaches recordings a bare walk cannot. Each kept recording carries a `match_reason` (\"title\" / \"source: …\" / \"speaker: …\") explaining why. Composes with project_id, limit, and cursor. Use it to find a recording by who is in it or what it's called, e.g. query: \"Andrew\"."
      },
      "project_id": {
        "type": "string",
        "description": "Optional project public ID — scope the list to a single project's recordings."
      },
      "limit": {
        "type": "integer",
        "description": "Rows per page. Default 20, which is also the model-context page ceiling (a larger value still pages at 20 with a next_cursor to continue; app hydration callers may receive up to 50 in one call)."
      },
      "cursor": {
        "type": "string",
        "description": "Opaque continuation returned as next_cursor by the previous page of this listing. Keep query and project_id identical when continuing; omit cursor to restart from the newest recordings. Never invent or edit a cursor."
      },
      "result_profile": {
        "type": "string",
        "enum": [
          "app",
          "model"
        ],
        "description": "Caller profile. \"app\" (widgets/apps) keeps full structured rows for hydration; \"model\" returns the complete row story as prose only. Omit to use the server default."
      }
    },
    "required": [],
    "additionalProperties": false
  },
  "annotations": {
    "readOnlyHint": true,
    "destructiveHint": false,
    "idempotentHint": true,
    "openWorldHint": false
  },
  "securitySchemes": [
    {
      "type": "noauth"
    }
  ],
  "_meta": {
    "securitySchemes": [
      {
        "type": "noauth"
      }
    ],
    "ui": {
      "visibility": [
        "model",
        "app"
      ]
    },
    "openai/widgetAccessible": true
  }
}

App descriptor

{
  "name": "recordings_list",
  "title": "List recordings",
  "description": "List standalone recordings (raw source media) — the drawer, not the editorial unit. Use for source/ingest/readiness asks (\"latest searchable session\", \"what uploads are ready?\") or to inspect and select recordings inside a project (project_id). For one Project reel whose useful footage spans separate recordings rather than one parent Episode, inspect selected recording pixels directly with recording_sample_frames when needed, then pass recording_id ranges directly to one episode_create without derived_from_id. Never create a throwaway Episode merely to inspect a Recording. `query` matches title, project, or who is in it (\"the latest with Andrew\" matches Andrew as a speaker). PAGINATED, newest first: one call returns up to 20 rows and says when more exist; continue with the returned next_cursor (same query/project_id) as `cursor`. One listing walk reaches at most the newest 50 recordings — anything older is reachable only by narrowing with query or project_id, and the result states which case you are in. Model rows carry both the exact tool-only recording_id and the episode-of-one episode_id, plus captured_at, searchability status/notes, and a compact indexed description. Prefer the newest `searchability_status:\"searchable\"` row for latest-session asks; episode_pending means the Recording may not have an Episode-of-one, not that direct source-time frame inspection must wait. Not for ordinary composition browsing (episodes_list) and not for searching what was SAID (transcript_search). Read-only.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "query": {
        "type": "string",
        "description": "Optional case-insensitive filter — returns only recordings whose TITLE, a composited SOURCE TITLE, or a diarized SPEAKER NAME contains this text (substring). The filter scans the FULL catalog before the newest-50 listing window applies, so it also reaches recordings a bare walk cannot. Each kept recording carries a `match_reason` (\"title\" / \"source: …\" / \"speaker: …\") explaining why. Composes with project_id, limit, and cursor. Use it to find a recording by who is in it or what it's called, e.g. query: \"Andrew\"."
      },
      "project_id": {
        "type": "string",
        "description": "Optional project public ID — scope the list to a single project's recordings."
      },
      "limit": {
        "type": "integer",
        "description": "Rows per page. Default 20, which is also the model-context page ceiling (a larger value still pages at 20 with a next_cursor to continue; app hydration callers may receive up to 50 in one call)."
      },
      "cursor": {
        "type": "string",
        "description": "Opaque continuation returned as next_cursor by the previous page of this listing. Keep query and project_id identical when continuing; omit cursor to restart from the newest recordings. Never invent or edit a cursor."
      },
      "result_profile": {
        "type": "string",
        "enum": [
          "app",
          "model"
        ],
        "description": "Caller profile. \"app\" (widgets/apps) keeps full structured rows for hydration; \"model\" returns the complete row story as prose only. Omit to use the server default."
      }
    },
    "required": [],
    "additionalProperties": false
  },
  "annotations": {
    "readOnlyHint": true,
    "destructiveHint": false,
    "idempotentHint": true,
    "openWorldHint": false
  },
  "securitySchemes": [
    {
      "type": "noauth"
    }
  ],
  "_meta": {
    "securitySchemes": [
      {
        "type": "noauth"
      }
    ],
    "ui": {
      "visibility": [
        "model",
        "app"
      ]
    },
    "openai/widgetAccessible": true
  }
}

Live Workspace descriptor

{
  "name": "recordings_list",
  "title": "List recordings",
  "description": "List standalone recordings (raw source media) — the drawer, not the editorial unit. Use for source/ingest/readiness asks (\"latest searchable session\", \"what uploads are ready?\") or to inspect and select recordings inside a project (project_id). For one Project reel whose useful footage spans separate recordings rather than one parent Episode, inspect selected recording pixels directly with recording_sample_frames when needed, then pass recording_id ranges directly to one episode_create without derived_from_id. Never create a throwaway Episode merely to inspect a Recording. `query` matches title, project, or who is in it (\"the latest with Andrew\" matches Andrew as a speaker). PAGINATED, newest first: one call returns up to 20 rows and says when more exist; continue with the returned next_cursor (same query/project_id) as `cursor`. One listing walk reaches at most the newest 50 recordings — anything older is reachable only by narrowing with query or project_id, and the result states which case you are in. Model rows carry both the exact tool-only recording_id and the episode-of-one episode_id, plus captured_at, searchability status/notes, and a compact indexed description. Prefer the newest `searchability_status:\"searchable\"` row for latest-session asks; episode_pending means the Recording may not have an Episode-of-one, not that direct source-time frame inspection must wait. Not for ordinary composition browsing (episodes_list) and not for searching what was SAID (transcript_search). Read-only.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "query": {
        "type": "string",
        "description": "Optional case-insensitive filter — returns only recordings whose TITLE, a composited SOURCE TITLE, or a diarized SPEAKER NAME contains this text (substring). The filter scans the FULL catalog before the newest-50 listing window applies, so it also reaches recordings a bare walk cannot. Each kept recording carries a `match_reason` (\"title\" / \"source: …\" / \"speaker: …\") explaining why. Composes with project_id, limit, and cursor. Use it to find a recording by who is in it or what it's called, e.g. query: \"Andrew\"."
      },
      "project_id": {
        "type": "string",
        "description": "Optional project public ID — scope the list to a single project's recordings."
      },
      "limit": {
        "type": "integer",
        "description": "Rows per page. Default 20, which is also the model-context page ceiling (a larger value still pages at 20 with a next_cursor to continue; app hydration callers may receive up to 50 in one call)."
      },
      "cursor": {
        "type": "string",
        "description": "Opaque continuation returned as next_cursor by the previous page of this listing. Keep query and project_id identical when continuing; omit cursor to restart from the newest recordings. Never invent or edit a cursor."
      },
      "result_profile": {
        "type": "string",
        "enum": [
          "app",
          "model"
        ],
        "description": "Caller profile. \"app\" (widgets/apps) keeps full structured rows for hydration; \"model\" returns the complete row story as prose only. Omit to use the server default."
      }
    },
    "required": [],
    "additionalProperties": false
  },
  "annotations": {
    "readOnlyHint": true,
    "destructiveHint": false,
    "idempotentHint": true,
    "openWorldHint": false
  },
  "securitySchemes": [
    {
      "type": "noauth"
    }
  ],
  "_meta": {
    "securitySchemes": [
      {
        "type": "noauth"
      }
    ]
  }
}

Errors

[]

Examples

[
  {
    "label": "List recent recordings",
    "input": {
      "limit": 10
    }
  },
  {
    "label": "Find a recording by speaker or title",
    "input": {
      "query": "Andrew"
    }
  },
  {
    "label": "List the recordings in a project",
    "input": {
      "project_id": "proj_..."
    }
  }
]