Troubleshooting generations
Keep the job ID returned by a generation request. It connects your request, provider status, output, and credit history so support can investigate the same generation.
A generation is taking longer than expected
Check the existing job with GET /api/v1/jobs/{id} or open Generation history in your dashboard. Pending or running means there is no confirmed final result yet. A temporary provider status error does not necessarily mean generation failed.
Wait between status checks and honor rate-limit responses. Submitting a new generation can create a separate charge. For Studio automation, reuse your original idempotency key when retrying the same run request.
A generation failed
Record the error code and message, model, settings, job ID, and approximate time. Check the job and your billing history for a credit refund before retrying. If a refund is missing, contact support with the job ID. A credit refund does not cancel a provider request or refund a card payment.
What to include when contacting support
- Job ID, or the full error response if no job was created.
- Your account email or organization ID.
- Model, resolution, duration, and when the request was submitted.
- What you expected and what happened instead.
Do not send API keys, authorization headers, passwords, or signed media URLs. Support has an internal diagnostic console for generation status and credit history.