Skip to main content

​
Track progress

Poll GET /v1/jobs/{job_id}?include_result=true every 30–60 seconds. Save the job ID returned at acceptance. Stop polling when job.status is succeeded, failed, or cancelled. Inspect job.error on failure. For batches, poll the batch and paginate its items rather than polling every job.

​
Cancel a job

Cancellation requests stop unfinished work; running workers may take time to stop. Poll until terminal. Completed work is not undone. Use the batch cancellation endpoint to stop a whole batch’s unfinished items.

​
Retry generation safely

Give each intended generation a unique Idempotency-Key (1–128 ASCII letters, digits, dots, underscores, colons or hyphens):
Retry a lost response with the same key and same JSON body. Once accepted, retries return the original job, with Idempotency-Replayed: true. A different body with the same key returns 409 idempotency_key_reused for Personal keys or 409 idempotency_conflict for Team keys. For Personal requests, concurrent or interrupted acceptance returns 409 request_in_progress with Retry-After: 5; retry the same key. If that persists, contact support with X-Request-ID rather than creating a new key and risking a duplicate job. Keys are scoped to your account and API deployment, and currently retained without automatic expiry. Recorded 4xx responses are replayed too; after correcting such a request, use a new key. Personal requests without a key create a new job each time; Team requests without a key are rejected. The 202 response includes Location: /v1/jobs/{id} and Retry-After: 30.

​
Completion notifications

Generation continues on the server when your client disconnects. Save the job_id to check the result later; you do not need to keep sending requests to keep generation running. Use polling every 30–60 seconds for all account scopes. Personal-key clients can alternatively keep one SSE connection for progress and completion; Team SSE currently returns the restriction described above. The API does not provide webhooks or a guaranteed completion email. Any account email notifications are supplementary; use the job status or SSE done event to drive your integration.

​
Streaming progress

SSE delivers a sequence of server events over one HTTP response, including progress and the terminal done event. It avoids repeated status requests. The currently deployed handler supports personal jobs; Team jobs must use polling until Team event-store integration is available.
Progress events include a sequence number. Reconnect with ?after=LAST_SEQUENCE to receive later events. Progress event types can vary; ignore types your client does not recognize. ping is a keepalive. done includes the terminal status and job_id, and closes the stream. Fetch the job with include_result=true to retrieve the asset after success.

​
Retry export without regeneration

If Blender authoring already succeeded and only export failed, this generation retry returns 409 export_retry_required. Use the saved-artifact recovery endpoint:
This Team-only operation returns 202 with a new job_id and Location. Poll that job and use its new asset_id for downloads. The original job, batch counts and artifacts remain unchanged. retry_of_job_id identifies the previous job and retry_mode is export. When known, failure_stage distinguishes blender, export, and parent failures. Recovery requires successful saved authoring, stopped workers, settled prior charges and an available authoring ZIP. It uses the original pinned release, not whichever exporter was most recently deployed. Retain that release’s worker app. Recovery cannot start Blender authoring or paid model/image calls. USD is exported again and original GLB bytes are retained. Extra formats, if configured, follow that original release’s export policy. Geometry, physics, joints and materials are reused; only the library record ID changes for the new result. This operation is not geometry repair or physics validation. Normal Team capacity and its configured billing policy still apply. Do not assume a successful export retry is free or that every policy creates a credit hold. Modal compute costs remain separate from customer pricing. Reuse the same idempotency key after a lost response. If this recovery also fails, its job can be passed to retry-export again; saved-authoring lineage remains tied to the original source.

​
Endpoint reference

Job status · Cancel a job · Retry export