View as Markdown

Default model tools

capture_session_mutate

Change a footage session

One footage-session change per call: add_recordings, camera_group_set, listening_reference_set, remove_recording, set_recording_start, or retry_analysis. add_recordings attaches existing project recordings (top-level recording_ids) and schedules incremental chronology analysis (bytes still on a device, or a later upload from another camera operator, arrive through upload_link_create with capture_session_id); membership is retry-safe and never creates unrelated top-level episodes. camera_group_set assigns the top-level recording_ids to one stable camera group without changing chronology, source identity, or coverage: it never advances the chronology revision, never returns the session to analysis, and never queues another solve (without camera_group_key it mints a new group; camera_label renames without becoming identity; camera identity carries its own generation, and expected_camera_generation is required so a concurrent camera edit cannot be silently overwritten; the receipt retains prior keys and labels so grouping reverses through the same action). listening_reference_set chooses the default speech-rich listening reference camera via top-level recording_id, or omits it to restore automatic selection; it never selects Program camera or audio, discards transcripts, or overrides speaker identity. remove_recording is a revision-guarded, audited exclusion of one recording (top-level recording_id) that preserves the original recording, membership history, observations, and solve receipts. set_recording_start is an audited placement correction setting one unresolved recording's start on the stable capture clock; its reason records the capture evidence or human instruction behind the placement. retry_analysis retries the same evidence revision after bounded automatic chronology attempts have failed; it applies when next_action.kind is retry_analysis, with one idempotency key per exact terminal failure generation. Subject references ride top-level: capture_session_id always; recording_id / recording_ids per action.

Surface: Default model and app

Product release eb8538f4a297f331d0399351c1e175bffcd212d2
Public contract commit 20688e11af473a363145ed880d5019c90a7ec1e2
Public contract SHA-256 87a94b85a84dc19245fc03f9b544c70d64bd9f5e08585dcccb38157cace5700d
Captured 2026-10-04T11:55:45.577Z. 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": "capture_session_mutate",
  "title": "Change a footage session",
  "description": "One footage-session change per call: add_recordings, camera_group_set, listening_reference_set, remove_recording, set_recording_start, or retry_analysis. add_recordings attaches existing project recordings (top-level recording_ids) and schedules incremental chronology analysis (bytes still on a device, or a later upload from another camera operator, arrive through upload_link_create with capture_session_id); membership is retry-safe and never creates unrelated top-level episodes. camera_group_set assigns the top-level recording_ids to one stable camera group without changing chronology, source identity, or coverage: it never advances the chronology revision, never returns the session to analysis, and never queues another solve (without camera_group_key it mints a new group; camera_label renames without becoming identity; camera identity carries its own generation, and expected_camera_generation is required so a concurrent camera edit cannot be silently overwritten; the receipt retains prior keys and labels so grouping reverses through the same action). listening_reference_set chooses the default speech-rich listening reference camera via top-level recording_id, or omits it to restore automatic selection; it never selects Program camera or audio, discards transcripts, or overrides speaker identity. remove_recording is a revision-guarded, audited exclusion of one recording (top-level recording_id) that preserves the original recording, membership history, observations, and solve receipts. set_recording_start is an audited placement correction setting one unresolved recording's start on the stable capture clock; its reason records the capture evidence or human instruction behind the placement. retry_analysis retries the same evidence revision after bounded automatic chronology attempts have failed; it applies when next_action.kind is retry_analysis, with one idempotency key per exact terminal failure generation. Subject references ride top-level: capture_session_id always; recording_id / recording_ids per action.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "capture_session_id": {
        "type": "string",
        "description": "Footage-session public ID."
      },
      "recording_id": {
        "type": "string",
        "description": "Recording public ID for listening_reference_set (optional; absent means automatic), remove_recording, and set_recording_start."
      },
      "recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "maxItems": 200,
        "description": "Recording public IDs for add_recordings and camera_group_set."
      },
      "expected_revision": {
        "type": "integer",
        "description": "Current footage-session revision. Required for every action except add_recordings; a stale revision is rejected."
      },
      "expected_camera_generation": {
        "type": "integer",
        "description": "Required for camera_group_set: the camera-identity generation this decision was made against, from manifest.camera_grouping.camera_generation. Camera identity has its own generation because grouping never advances the footage-session revision, so expected_revision cannot detect a concurrent camera decision. Omitting it is rejected; a stale one is rejected."
      },
      "idempotency_key": {
        "type": "string",
        "description": "Caller-stable key for safe retry of this exact change."
      },
      "action": {
        "type": "object",
        "description": "One footage-session action. Subject references (capture_session_id, recording_id, recording_ids) ride top-level.",
        "oneOf": [
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "add_recordings"
                ]
              },
              "relative_paths": {
                "type": "object",
                "description": "Optional map from recording ID to a safe device/folder-relative path retained as ingest context."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "camera_group_set"
                ]
              },
              "camera_group_key": {
                "type": "string",
                "description": "Optional existing group key from this session; a new group is created when omitted."
              },
              "camera_label": {
                "type": "string",
                "description": "Optional human label. Renaming never changes group identity."
              },
              "reason": {
                "type": "string",
                "description": "Required evidence or explicit human instruction supporting the grouping."
              }
            },
            "required": [
              "type",
              "reason"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "listening_reference_set"
                ]
              },
              "reason": {
                "type": "string",
                "description": "Optional concise reason retained with the preference receipt."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "remove_recording"
                ]
              },
              "reason": {
                "type": "string",
                "description": "Optional concise reason retained in the audit receipt."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "set_recording_start"
                ]
              },
              "capture_start_seconds": {
                "type": "number",
                "description": "Recording start on the session's stable capture clock. May be negative when footage predates the coordinate origin."
              },
              "reason": {
                "type": "string",
                "description": "Required concise rationale citing the capture evidence or explicit human instruction that establishes this placement."
              }
            },
            "required": [
              "type",
              "capture_start_seconds",
              "reason"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "retry_analysis"
                ]
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          }
        ]
      }
    },
    "required": [
      "capture_session_id",
      "idempotency_key",
      "action"
    ],
    "additionalProperties": false
  },
  "annotations": {
    "title": "Change a footage session",
    "readOnlyHint": false,
    "destructiveHint": false,
    "idempotentHint": true,
    "openWorldHint": false
  },
  "securitySchemes": [
    {
      "type": "noauth"
    }
  ],
  "_meta": {
    "securitySchemes": [
      {
        "type": "noauth"
      }
    ],
    "ui": {
      "visibility": [
        "model",
        "app"
      ]
    },
    "openai/widgetAccessible": true
  }
}

App descriptor

{
  "name": "capture_session_mutate",
  "title": "Change a footage session",
  "description": "One footage-session change per call: add_recordings, camera_group_set, listening_reference_set, remove_recording, set_recording_start, or retry_analysis. add_recordings attaches existing project recordings (top-level recording_ids) and schedules incremental chronology analysis (bytes still on a device, or a later upload from another camera operator, arrive through upload_link_create with capture_session_id); membership is retry-safe and never creates unrelated top-level episodes. camera_group_set assigns the top-level recording_ids to one stable camera group without changing chronology, source identity, or coverage: it never advances the chronology revision, never returns the session to analysis, and never queues another solve (without camera_group_key it mints a new group; camera_label renames without becoming identity; camera identity carries its own generation, and expected_camera_generation is required so a concurrent camera edit cannot be silently overwritten; the receipt retains prior keys and labels so grouping reverses through the same action). listening_reference_set chooses the default speech-rich listening reference camera via top-level recording_id, or omits it to restore automatic selection; it never selects Program camera or audio, discards transcripts, or overrides speaker identity. remove_recording is a revision-guarded, audited exclusion of one recording (top-level recording_id) that preserves the original recording, membership history, observations, and solve receipts. set_recording_start is an audited placement correction setting one unresolved recording's start on the stable capture clock; its reason records the capture evidence or human instruction behind the placement. retry_analysis retries the same evidence revision after bounded automatic chronology attempts have failed; it applies when next_action.kind is retry_analysis, with one idempotency key per exact terminal failure generation. Subject references ride top-level: capture_session_id always; recording_id / recording_ids per action.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "capture_session_id": {
        "type": "string",
        "description": "Footage-session public ID."
      },
      "recording_id": {
        "type": "string",
        "description": "Recording public ID for listening_reference_set (optional; absent means automatic), remove_recording, and set_recording_start."
      },
      "recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "maxItems": 200,
        "description": "Recording public IDs for add_recordings and camera_group_set."
      },
      "expected_revision": {
        "type": "integer",
        "description": "Current footage-session revision. Required for every action except add_recordings; a stale revision is rejected."
      },
      "expected_camera_generation": {
        "type": "integer",
        "description": "Required for camera_group_set: the camera-identity generation this decision was made against, from manifest.camera_grouping.camera_generation. Camera identity has its own generation because grouping never advances the footage-session revision, so expected_revision cannot detect a concurrent camera decision. Omitting it is rejected; a stale one is rejected."
      },
      "idempotency_key": {
        "type": "string",
        "description": "Caller-stable key for safe retry of this exact change."
      },
      "action": {
        "type": "object",
        "description": "One footage-session action. Subject references (capture_session_id, recording_id, recording_ids) ride top-level.",
        "oneOf": [
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "add_recordings"
                ]
              },
              "relative_paths": {
                "type": "object",
                "description": "Optional map from recording ID to a safe device/folder-relative path retained as ingest context."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "camera_group_set"
                ]
              },
              "camera_group_key": {
                "type": "string",
                "description": "Optional existing group key from this session; a new group is created when omitted."
              },
              "camera_label": {
                "type": "string",
                "description": "Optional human label. Renaming never changes group identity."
              },
              "reason": {
                "type": "string",
                "description": "Required evidence or explicit human instruction supporting the grouping."
              }
            },
            "required": [
              "type",
              "reason"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "listening_reference_set"
                ]
              },
              "reason": {
                "type": "string",
                "description": "Optional concise reason retained with the preference receipt."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "remove_recording"
                ]
              },
              "reason": {
                "type": "string",
                "description": "Optional concise reason retained in the audit receipt."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "set_recording_start"
                ]
              },
              "capture_start_seconds": {
                "type": "number",
                "description": "Recording start on the session's stable capture clock. May be negative when footage predates the coordinate origin."
              },
              "reason": {
                "type": "string",
                "description": "Required concise rationale citing the capture evidence or explicit human instruction that establishes this placement."
              }
            },
            "required": [
              "type",
              "capture_start_seconds",
              "reason"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "retry_analysis"
                ]
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          }
        ]
      }
    },
    "required": [
      "capture_session_id",
      "idempotency_key",
      "action"
    ],
    "additionalProperties": false
  },
  "annotations": {
    "title": "Change a footage session",
    "readOnlyHint": false,
    "destructiveHint": false,
    "idempotentHint": true,
    "openWorldHint": false
  },
  "securitySchemes": [
    {
      "type": "noauth"
    }
  ],
  "_meta": {
    "securitySchemes": [
      {
        "type": "noauth"
      }
    ],
    "ui": {
      "visibility": [
        "model",
        "app"
      ]
    },
    "openai/widgetAccessible": true
  },
  "outputSchema": {
    "type": "object",
    "properties": {
      "capture_session_id": {
        "type": "string"
      },
      "episode_id": {
        "type": "string"
      },
      "revision": {
        "type": "integer"
      },
      "status": {
        "type": "string"
      },
      "already_applied": {
        "type": "boolean"
      },
      "analysis_enqueued": {
        "type": "boolean"
      },
      "dispatch_status": {
        "type": "string",
        "enum": [
          "dispatching",
          "queued",
          "observed",
          "enqueue_failed"
        ],
        "description": "retry_analysis only: durable queue-delivery state for this retry receipt."
      },
      "recording_id": {
        "type": [
          "string",
          "null"
        ]
      },
      "recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        }
      },
      "added_recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        }
      },
      "camera_group_key": {
        "type": [
          "string",
          "null"
        ]
      },
      "camera_label": {
        "type": [
          "string",
          "null"
        ]
      },
      "camera_generation": {
        "type": "integer",
        "description": "camera_group_set only: the footage session's camera-identity generation after this call, which the next camera edit takes as expected_camera_generation."
      },
      "prior_memberships": {
        "type": "array",
        "items": {
          "type": "object"
        }
      },
      "removed_membership_id": {
        "type": "string"
      },
      "source_preserved": {
        "type": "boolean"
      },
      "correction_id": {
        "type": "string"
      },
      "capture_start_seconds": {
        "type": "number"
      },
      "selection_basis": {
        "type": "string",
        "enum": [
          "automatic",
          "user_selected"
        ]
      },
      "audit_receipt_id": {
        "type": "string"
      },
      "manifest": {
        "type": "object"
      },
      "next_action": {
        "type": "object"
      }
    },
    "required": [],
    "additionalProperties": false
  }
}

Live Workspace descriptor

{
  "name": "capture_session_mutate",
  "title": "Change a footage session",
  "description": "One footage-session change per call: add_recordings, camera_group_set, listening_reference_set, remove_recording, set_recording_start, or retry_analysis. add_recordings attaches existing project recordings (top-level recording_ids) and schedules incremental chronology analysis (bytes still on a device, or a later upload from another camera operator, arrive through upload_link_create with capture_session_id); membership is retry-safe and never creates unrelated top-level episodes. camera_group_set assigns the top-level recording_ids to one stable camera group without changing chronology, source identity, or coverage: it never advances the chronology revision, never returns the session to analysis, and never queues another solve (without camera_group_key it mints a new group; camera_label renames without becoming identity; camera identity carries its own generation, and expected_camera_generation is required so a concurrent camera edit cannot be silently overwritten; the receipt retains prior keys and labels so grouping reverses through the same action). listening_reference_set chooses the default speech-rich listening reference camera via top-level recording_id, or omits it to restore automatic selection; it never selects Program camera or audio, discards transcripts, or overrides speaker identity. remove_recording is a revision-guarded, audited exclusion of one recording (top-level recording_id) that preserves the original recording, membership history, observations, and solve receipts. set_recording_start is an audited placement correction setting one unresolved recording's start on the stable capture clock; its reason records the capture evidence or human instruction behind the placement. retry_analysis retries the same evidence revision after bounded automatic chronology attempts have failed; it applies when next_action.kind is retry_analysis, with one idempotency key per exact terminal failure generation. Subject references ride top-level: capture_session_id always; recording_id / recording_ids per action.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "capture_session_id": {
        "type": "string",
        "description": "Footage-session public ID."
      },
      "recording_id": {
        "type": "string",
        "description": "Recording public ID for listening_reference_set (optional; absent means automatic), remove_recording, and set_recording_start."
      },
      "recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "maxItems": 200,
        "description": "Recording public IDs for add_recordings and camera_group_set."
      },
      "expected_revision": {
        "type": "integer",
        "description": "Current footage-session revision. Required for every action except add_recordings; a stale revision is rejected."
      },
      "expected_camera_generation": {
        "type": "integer",
        "description": "Required for camera_group_set: the camera-identity generation this decision was made against, from manifest.camera_grouping.camera_generation. Camera identity has its own generation because grouping never advances the footage-session revision, so expected_revision cannot detect a concurrent camera decision. Omitting it is rejected; a stale one is rejected."
      },
      "idempotency_key": {
        "type": "string",
        "description": "Caller-stable key for safe retry of this exact change."
      },
      "action": {
        "type": "object",
        "description": "One footage-session action. Subject references (capture_session_id, recording_id, recording_ids) ride top-level.",
        "oneOf": [
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "add_recordings"
                ]
              },
              "relative_paths": {
                "type": "object",
                "description": "Optional map from recording ID to a safe device/folder-relative path retained as ingest context."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "camera_group_set"
                ]
              },
              "camera_group_key": {
                "type": "string",
                "description": "Optional existing group key from this session; a new group is created when omitted."
              },
              "camera_label": {
                "type": "string",
                "description": "Optional human label. Renaming never changes group identity."
              },
              "reason": {
                "type": "string",
                "description": "Required evidence or explicit human instruction supporting the grouping."
              }
            },
            "required": [
              "type",
              "reason"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "listening_reference_set"
                ]
              },
              "reason": {
                "type": "string",
                "description": "Optional concise reason retained with the preference receipt."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "remove_recording"
                ]
              },
              "reason": {
                "type": "string",
                "description": "Optional concise reason retained in the audit receipt."
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "set_recording_start"
                ]
              },
              "capture_start_seconds": {
                "type": "number",
                "description": "Recording start on the session's stable capture clock. May be negative when footage predates the coordinate origin."
              },
              "reason": {
                "type": "string",
                "description": "Required concise rationale citing the capture evidence or explicit human instruction that establishes this placement."
              }
            },
            "required": [
              "type",
              "capture_start_seconds",
              "reason"
            ],
            "additionalProperties": false
          },
          {
            "type": "object",
            "properties": {
              "type": {
                "type": "string",
                "enum": [
                  "retry_analysis"
                ]
              }
            },
            "required": [
              "type"
            ],
            "additionalProperties": false
          }
        ]
      }
    },
    "required": [
      "capture_session_id",
      "idempotency_key",
      "action"
    ],
    "additionalProperties": false
  },
  "annotations": {
    "title": "Change a footage session",
    "readOnlyHint": false,
    "destructiveHint": false,
    "idempotentHint": true,
    "openWorldHint": false
  },
  "securitySchemes": [
    {
      "type": "noauth"
    }
  ],
  "_meta": {
    "securitySchemes": [
      {
        "type": "noauth"
      }
    ]
  },
  "outputSchema": {
    "type": "object",
    "properties": {
      "capture_session_id": {
        "type": "string"
      },
      "episode_id": {
        "type": "string"
      },
      "revision": {
        "type": "integer"
      },
      "status": {
        "type": "string"
      },
      "already_applied": {
        "type": "boolean"
      },
      "analysis_enqueued": {
        "type": "boolean"
      },
      "dispatch_status": {
        "type": "string",
        "enum": [
          "dispatching",
          "queued",
          "observed",
          "enqueue_failed"
        ],
        "description": "retry_analysis only: durable queue-delivery state for this retry receipt."
      },
      "recording_id": {
        "type": [
          "string",
          "null"
        ]
      },
      "recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        }
      },
      "added_recording_ids": {
        "type": "array",
        "items": {
          "type": "string"
        }
      },
      "camera_group_key": {
        "type": [
          "string",
          "null"
        ]
      },
      "camera_label": {
        "type": [
          "string",
          "null"
        ]
      },
      "camera_generation": {
        "type": "integer",
        "description": "camera_group_set only: the footage session's camera-identity generation after this call, which the next camera edit takes as expected_camera_generation."
      },
      "prior_memberships": {
        "type": "array",
        "items": {
          "type": "object"
        }
      },
      "removed_membership_id": {
        "type": "string"
      },
      "source_preserved": {
        "type": "boolean"
      },
      "correction_id": {
        "type": "string"
      },
      "capture_start_seconds": {
        "type": "number"
      },
      "selection_basis": {
        "type": "string",
        "enum": [
          "automatic",
          "user_selected"
        ]
      },
      "audit_receipt_id": {
        "type": "string"
      },
      "manifest": {
        "type": "object"
      },
      "next_action": {
        "type": "object"
      }
    },
    "required": [],
    "additionalProperties": false
  }
}

Errors

[
  "capture_session_not_found",
  "capture_session_not_failed",
  "recording_not_found",
  "stale_capture_session_revision",
  "idempotency_conflict",
  "invalid_input"
]

Examples

[]