> ## Documentation Index > Fetch the complete documentation index at: https://docs.rundesert.com/llms.txt > Use this file to discover all available pages before exploring further. # Get a run's result > What the automation produced. The shape of `result` is the automation's own: two automations answering different questions have no reason to agree on one, so nothing here reshapes it. An unfinished or failed run answers **409**, not a null result. That keeps "there is nothing yet" apart from "this automation returned null", which is a real answer some of them give. ## OpenAPI ````yaml /openapi.json get /api/automations/{automationId}/runs/{runId}/result openapi: 3.1.0 info: title: Workspace API version: 1.0.0 description: >- Every workspace answers these endpoints. Nothing is registered and nothing is switched on: an automation that appears in your Automations tab is already readable here. The base URL is your workspace's own. Ask the agent for it, or find it under Settings → API on the workspace. It is stable and never changes. Every call carries your workspace's API key, as `Authorization: Bearer ` or in the `x-desert-api-key` header. The key is on the same settings page as the URL, and generating a new one there revokes the old one immediately. servers: - url: https://www.rundesert.com/p/{workspaceId} description: Your workspace's public API variables: workspaceId: default: px_XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX description: >- The id in your workspace's public URL. It is the address, not the credential — the API key is what authorises the call. security: - bearerAuth: [] - apiKeyHeader: [] tags: - name: Automations description: What this workspace can do. - name: Runs description: What it has done, and what came out. - name: Files description: Downloading what an automation produced. - name: Custom endpoints description: Anything else your agent wrote. paths: /api/automations/{automationId}/runs/{runId}/result: get: tags: - Runs summary: Get a run's result description: >- What the automation produced. The shape of `result` is the automation's own: two automations answering different questions have no reason to agree on one, so nothing here reshapes it. An unfinished or failed run answers **409**, not a null result. That keeps "there is nothing yet" apart from "this automation returned null", which is a real answer some of them give. operationId: getRunResult parameters: - $ref: '#/components/parameters/AutomationId' - $ref: '#/components/parameters/RunId' responses: '200': description: The run finished and this is what it produced. content: application/json: schema: $ref: '#/components/schemas/RunResult' example: runId: 1786945498854-80788d automation: invoice-extract status: done duration: 10277 result: |- **Supplier** — Acme Supply Co. **Amount** — $1,240.00 **Due** — 2026-09-01 '401': $ref: '#/components/responses/Unauthorized' '404': $ref: '#/components/responses/NotFound' '409': description: >- The run has not finished, or it failed. Poll the run endpoint until `status` is `done`. content: application/json: schema: $ref: '#/components/schemas/NotReady' examples: stillRunning: summary: Still running value: error: the run has not finished status: running duration: null poll: /api/automations/invoice-extract/runs/1786945498854-80788d failed: summary: Failed value: error: the run failed status: error detail: The invoice could not be read. duration: 4120 '429': $ref: '#/components/responses/RateLimited' components: parameters: AutomationId: name: automationId in: path required: true description: The automation's id, as returned by `GET /api/automations`. schema: type: string example: invoice-extract RunId: name: runId in: path required: true description: >- A run id. A run is only readable through the automation that produced it. schema: type: string example: 1786945498854-80788d schemas: RunResult: type: object properties: runId: type: string automation: type: string status: $ref: '#/components/schemas/RunStatus' duration: type: integer description: Milliseconds. result: description: >- Whatever the automation returned, unchanged. Often markdown, often an object. Ask the agent what a given automation returns, or read one finished run. oneOf: - type: string - type: object - type: array - type: number - type: boolean - type: 'null' NotReady: type: object properties: error: type: string status: $ref: '#/components/schemas/RunStatus' detail: type: string description: Present when the run failed. duration: type: - integer - 'null' poll: type: string description: 'Present when the run is still going: the path to check.' RunStatus: type: string enum: - running - done - error description: >- `running` means it has not finished. `error` means it finished badly, and `error` on the run says why. Error: type: object properties: error: type: string detail: type: string responses: Unauthorized: description: >- The API key was missing or wrong. Check Settings → API for the current one — generating a new key there stops the old one working immediately. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: unauthorized NotFound: description: No such automation, or no such run under it. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: No such automation RateLimited: description: >- 600 requests a minute across the whole workspace, shared with anyone loading its pages. Wait and try again. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: rate limited securitySchemes: bearerAuth: type: http scheme: bearer description: >- Your workspace's API key, from Settings → API. Sent as `Authorization: Bearer dk_...`. apiKeyHeader: type: apiKey in: header name: x-desert-api-key description: >- The same key, for callers whose `Authorization` header is already spoken for by their own auth. Send one form or the other, not both. ````