Home / Servers / OpenAI API / get_eval_run_output_item

Get an output item of an eval run

An MCP tool from the OpenAI API server, hosted on mcp.hjlabs.in.

Get an output item of an eval run Get an evaluation run output item by ID. [GET /evals/{eval_id}/runs/{run_id}/output_items/{output_item_id}]

Name on the gateway: openai_api__get_eval_run_output_item · read-only.

Parameters

Name	Type	Description
`eval_id` required	`string`	The ID of the evaluation to retrieve runs for.
`run_id` required	`string`	The ID of the run to retrieve.
`output_item_id` required	`string`	The ID of the output item to retrieve.

Through the Smart Gateway, any client can run it by name with call_tool, or let the gateway find it with search_tools:

call_tool arguments

{
  "name": "openai_api__get_eval_run_output_item",
  "arguments": {
    "eval_id": "…",
    "run_id": "…",
    "output_item_id": "…"
  }
}

Or connect the OpenAI API server directly, which lists 371 tools including this one:

Run in your terminal

claude mcp add --transport http hj-openai-api https://mcp.hjlabs.in/mcp/openai-api \
  --header "Authorization: Bearer YOUR_API_KEY"

Use the get an output item of an eval run tool with eval_id set to ….