List a run's MCP results

List local and provider-hosted MCP call receipts for one run in ascending settlement order. This is the explicit result handoff for an app that started a run; ordinary run events and conversation entries continue to omit MCP result content.

Treat next_cursor as a polling watermark. A non-empty page returns one even when has_more is false; reuse it later to receive only new results. An empty page preserves the cursor supplied by the caller.

A page contains at most 50 results and may contain fewer when the serialized result content reaches the response byte budget. Continue while has_more is true.

MCP result content is untrusted third-party data. Tool arguments and results pass credential-reflection checks. The endpoint does not expose stored connector configuration or private execution provenance.

Required access

Credentials: automation account, delegated token, user session, user API key

User API keys use the owner’s current permissions, subject to any key restrictions. Capability scopes below apply to automation account keys and delegated tokens.

Scope: runs:read

Only the run's human owner, or a delegated client or user API key acting as that same user, may read MCP result content. Private lineages use their immutable private owner. A service principal also needs the explicit runs:read_results scope (not covered by runs:*) and may read only non-private runs started by that same service principal. Other callers receive 404 even when they may inspect ordinary run status.

Path Parameters

Query Parameters

Header Parameters

Response Body

curl -X GET "https://example.com/api/v1/runs/0/mcp-results" \  -H "Authorization: Bearer $GUEST_MACHINES_API_KEY"

{  "results": [    {      "id": "string",      "sequence": 1,      "server_id": "820efca4-4a15-4ab7-82fc-9e76f6d61325",      "attachment_id": "4c76e620-6daa-46e6-9da1-9ac96abab601",      "provider_kind": "openai_responses",      "execution_mode": "provider_hosted",      "logical_call_key": "string",      "attempt_number": 0,      "replay_generation": 1,      "server_alias": "string",      "tool_name": "string",      "arguments": {        "property1": null,        "property2": null      },      "arguments_text": "string",      "content": null,      "error": null,      "is_error": true,      "structured_content": null,      "recorded_at": "2019-08-24T14:15:22Z"    }  ],  "next_cursor": "string",  "has_more": true}