Get Run Status

Get the current status of an agent run. Use this to poll progress for long-running agent executions.

Request

GET /api/v1/agent/run/{runId}/status

Path Parameters

Parameter	Type	Description
`runId`	`string`	Agent run ID

Headers

Header	Required	Description
`Authorization`	Yes	API key in `Bearer snorbe_...` format

Response

{
  "id": "clxyz...",
  "status": "completed",
  "createdAt": "2026-04-16T09:36:01.351Z",
  "updatedAt": "2026-04-16T09:36:08.352Z",
  "pendingPlanDraft": false,
  "pendingReportDraft": false,
  "pendingMatrixDraft": false,
  "browseState": {
    "isBrowsing": true,
    "askHumanQuestion": "Which menu should I open after logging in?"
  },
  "skillState": {
    "isRunningSkill": true,
    "skillName": "patent-search",
    "pendingSecretKeys": ["PATENT_API_KEY"]
  }
}

Response Fields

Field	Type	Description
`id`	`string`	Agent run ID
`status`	`string`	Status (`"completed"`, `"running"`, `"error"`, etc.)
`createdAt`	`string`	Creation timestamp (ISO 8601)
`updatedAt`	`string`	Last update timestamp (ISO 8601)
`pendingPlanDraft`	`boolean`	Whether a plan draft is pending review
`pendingReportDraft`	`boolean`	Whether a report draft is pending review
`pendingMatrixDraft`	`boolean`	Whether a matrix draft is pending review
`browseState`	`object`	Returned only while the browse tool is active
`browseState.isBrowsing`	`boolean`	Whether browser automation is still active
`browseState.askHumanQuestion`	`string`	Question waiting for a human answer during browsing. Omitted when absent
`skillState`	`object`	Returned while a skill is running or waiting for secrets
`skillState.isRunningSkill`	`boolean`	Whether the skill session is active
`skillState.skillName`	`string`	Skill name that is running or requesting secrets
`skillState.pendingSecretKeys`	`string[]`	Secret key names waiting to be registered

When pendingPlanDraft, pendingReportDraft, or pendingMatrixDraft is true, you need to complete the review using the corresponding HITL endpoint. After that, resume execution via the streaming resume endpoint.

If browseState.askHumanQuestion is present, this is not a plan review. It is a browser automation question. Send the answer to /browser/answer-question, not to /agent/run/{runId}/.... The request needs the browse-start.payload.websocketInfo.session_id value captured from the active SSE stream.

If skillState.pendingSecretKeys has values, the skill is waiting for secrets. Register each key through /secret so the waiting skill execution can continue.

Examples

curl "https://app.snorbe.deskrex.ai/api/v1/agent/run/clxyz123/status" \
  -H "Authorization: Bearer snorbe_your_api_key_here"

import requests
import time

run_id = "clxyz123"

while True:
    resp = requests.get(
        f"https://app.snorbe.deskrex.ai/api/v1/agent/run/{run_id}/status",
        headers={"Authorization": "Bearer snorbe_your_api_key_here"},
    )
    data = resp.json()
    print(f"Status: {data['status']}")

    if data["status"] == "completed":
        print("Done!")
        break
    if data["status"] == "error":
        print(f"Error occurred")
        break

    time.sleep(5)

Error Responses

HTTP Status	Description
401	Invalid API key
400	Invalid `runId` format