When you press Generate, Rhinobird queues work behind the scenes and eventually saves outputs to your account. Most of the time the flow is simple: the tool page shows progress, then History lists a preview you can download. Sometimes the job sits in queue longer than expected, appears to freeze mid-run, or ends in a failed state with no file. This guide walks through those cases so you can recover without guessing—and so you know when credit returns apply versus when consumption may stand under membership rules.
Confirm you are signed in on the right account
Generation and History are tied to the account that started the job. If the page asks you to sign in again, complete Login / Sign up using Continue with Google or Continue with Email, then return to the tool and check History. An empty list often means you are on a different email than the one that pressed Generate, not that the platform lost your work globally.
Understand queued vs in progress vs failed
- Queued — The job is waiting for capacity. Peak hours and long video renders extend wait time. Stay on the tab or return later; do not assume failure in the first minute.
- In progress — The model is working. Closing the browser usually does not cancel server-side work, but session expiry can make the UI look unresponsive until you refresh.
- Failed — The task page or History shows an error or missing deliverable. Read the message before retrying; user-side input problems need fixes, not identical retries.
- Completed — Preview and download appear in History. If the tool page still spins, open History directly; the card may already be ready.
What to do when a job seems stuck
- Note the time you pressed Generate and the credit cost shown on the button.
- Wait a reasonable window for your media type—single images often finish sooner than multi-scene video.
- Refresh the tool page once, then open History and refresh there.
- If status is still ambiguous, avoid launching ten duplicate jobs; one stuck job plus nine copies wastes credits.
- When the UI offers cancel before execution begins, use it only if you are sure you do not want the run; cancellation timing affects whether credits return under platform rules.
Long renders are normal for high-duration video and large image sets. Check that your network is stable; intermittent connectivity can delay status updates without meaning the backend stopped.
When failed tasks may return credits
Under Rhinobird membership rules, pre-deducted credits or basic benefit usage may be returned when an AI task cannot complete because of system error, timeout, lack of results, or cancellation before execution begins. Returns are not automatic for every disappointment—for example, when the model produced an output you dislike but the job technically completed, or when failure traces to invalid uploads or policy issues on your side. Always compare the job record in History with credit transaction details for the same timestamp before you retry at full cost.
Retry safely after a real failure
After you confirm a return—or accept that consumption stands—fix the underlying issue before another Generate:
- Tighten Creative Requirements so product identity and mandatory copy are explicit.
- Replace blurry or cropped product photos; use preparation tools if labels must be readable.
- Lower duration or batch size for a proof run, then scale up once History shows acceptable quality.
- Re-read the cost on Generate; pricing reflects model and settings and may differ from the failed attempt.
When to contact support
Escalate when History shows failed or empty delivery, credits did not return within a reasonable processing window, or the same error repeats after you fixed inputs. Gather your account email, approximate job time, tool name, and screenshots of the task page. Reach support at 13378452864 or support@rhinobird.ai. For broader terms on AI tasks and credits, see Service Agreement and company information on About Us.
Patience plus one disciplined retry beats spamming Generate. History and credit details tell you whether the platform owes you a return or whether your next action is to improve inputs and brief text.