Get Response

GET

responses

{response_id}

curl https://gateway.bud.studio/v1/responses/resp_abc123 \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "x-model-name: gpt-4"

{
  "id": "resp_abc123",
  "object": "response",
  "created_at": 1699000000,
  "model": "gpt-4",
  "status": "completed",
  "conversation": { "id": "conv_9f2c1d7b4a6e08f35c1d2b7e4a9f6038" },
  "previous_response_id": null,
  "output": [
    {
      "id": "msg_xyz",
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "content": [
        { "type": "output_text", "text": "The capital of France is Paris.", "annotations": [] }
      ]
    }
  ],
  "usage": {
    "input_tokens": 15,
    "output_tokens": 8,
    "total_tokens": 23
  },
  "metadata": null
}

curl https://gateway.bud.studio/v1/responses/resp_abc123 \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "x-model-name: gpt-4"

{
  "id": "resp_abc123",
  "object": "response",
  "created_at": 1699000000,
  "model": "gpt-4",
  "status": "completed",
  "conversation": { "id": "conv_9f2c1d7b4a6e08f35c1d2b7e4a9f6038" },
  "previous_response_id": null,
  "output": [
    {
      "id": "msg_xyz",
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "content": [
        { "type": "output_text", "text": "The capital of France is Paris.", "annotations": [] }
      ]
    }
  ],
  "usage": {
    "input_tokens": 15,
    "output_tokens": 8,
    "total_tokens": 23
  },
  "metadata": null
}

Headers

Parameter	Type	Required	Description
Authorization	string	Yes	Bearer authentication header
x-model-name	string	No	Model name for routing. Default: gpt-4-responses

Path Parameters

Parameter	Type	Required	Description
response_id	string	Yes	The ID of the response to retrieve

The id resolves from the moment the run starts

A response id is retrievable for the whole life of the run, not only once it finishes. A GET issued while the run is still working returns the envelope with status: "in_progress" and an empty output — for a background run, for a streaming run, and for an ordinary synchronous one.

{
  "id": "resp_abc123",
  "object": "response",
  "status": "in_progress",
  "created_at": 1699000000,
  "conversation": { "id": "conv_9f2c…" },
  "output": [],
  "usage": null
}

That makes the id you receive in response.created (the first frame of a stream) usable immediately — to poll, to show the run in an operator view, or to cancel it.

A paused approval also reports status: "in_progress"; what identifies it is an mcp_approval_request item in output. See Governance outcomes.

When a response stops being retrievable

Cause	Result
The request set `store: false`	`404` once the answer has been returned — see what that guarantees.
The response was deleted	`404`. Deleting one turn does not remove the rest of its thread.
Retention window elapsed	`404`. Responses in a terminal state (`completed`, `failed`, `incomplete`, `cancelled`) are deleted after `RUN_RETENTION_DAYS` — 30 days by default, configurable per installation.

Runs that are not terminal — including one paused awaiting a human approval — are never collected by retention at any age, so a pending approval cannot expire out from under the client that is waiting on it.

Supported Providers

GET /v1/responses/{id} is provider-agnostic — every response Bud stores is retrievable this way, whichever model produced it.

Retrieval is scoped to the API key that created the response: another key asking for the same id gets 404, never another tenant’s content.

curl https://gateway.bud.studio/v1/responses/resp_abc123 \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "x-model-name: gpt-4"

{
  "id": "resp_abc123",
  "object": "response",
  "created_at": 1699000000,
  "model": "gpt-4",
  "status": "completed",
  "conversation": { "id": "conv_9f2c1d7b4a6e08f35c1d2b7e4a9f6038" },
  "previous_response_id": null,
  "output": [
    {
      "id": "msg_xyz",
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "content": [
        { "type": "output_text", "text": "The capital of France is Paris.", "annotations": [] }
      ]
    }
  ],
  "usage": {
    "input_tokens": 15,
    "output_tokens": 8,
    "total_tokens": 23
  },
  "metadata": null
}

Introduction

Gateway APIs

Python SDK & CLI

Headers

Path Parameters

The id resolves from the moment the run starts

When a response stops being retrievable

Supported Providers

​Headers

​Path Parameters

​The id resolves from the moment the run starts

​When a response stops being retrievable

​Supported Providers

Headers

Path Parameters

The id resolves from the moment the run starts

When a response stops being retrievable

Supported Providers