
Show the task state, elapsed waiting time and last successful status check instead of inventing a completion percentage. Imagild's current API does not provide percentage progress, queue position or a reliable finish-time estimate. An honest pending indicator can still explain what happened and what the user can do next.
Key Takeaways
- Separate your application's local waiting state from the remote task's reported status.
- Label elapsed time as waiting time, not generation completion.
- A successful generation still needs authenticated download and media review.
Name the state you actually know
Before submission, an item may be waiting in your application's own queue. After HTTP acceptance, it has a task ID and a remote status. Those are different facts, and using the same “queued” label for both without explanation hides useful information.
The Imagild API guide describes asynchronous submission and task reads. Your interface can show “Waiting to submit” before dispatch, then display the returned state after acceptance. Do not assign a provider queue number based on the order of rows in your own database.
| Evidence | Useful message | Avoid claiming |
|---|---|---|
| Local item has not been sent | Waiting in this application's queue | Provider position 3 |
| Task reports processing | Generating; last checked recently | 80% complete |
| Status read receives 429 | Updates paused; next check scheduled | Generation failed |
| Task reports unknown | Outcome needs reconciliation | Automatically refunded |
These are proposed interface messages, not additional API status values.
Use an indeterminate indicator accessibly
A subtle motion indicator can reassure a user that the interface is active without pretending to measure model progress. Keep the task label and recovery controls visible; do not replace the entire card with an endless animation.
For an HTML interface, this is a small illustrative pattern:
<section aria-busy="true" aria-describedby="generation-status">
<progress aria-label="Video generation"></progress>
<p id="generation-status" role="status">
Generating. Last successful status check: 12 seconds ago.
</p>
</section>
There is deliberately no numeric value. MDN's progressbar guidance explains the distinction between determinate and indeterminate progress. Update the text from real application observations rather than copying the example's twelve-second value.
Distinguish elapsed time from fresh evidence
The returned created_at describes the task record's creation time. A clock based on it can show elapsed waiting time, but it does not establish how long a GPU has been executing or how much work remains.
Store the time of your last successful status read separately. If polling fails, retain the last known remote state and show that it is stale. A network disconnect should not overwrite a known task with “failed.” A refresh should recover the saved task ID rather than create another generation.
The failed-job checks help explain when an error calls for investigation instead of another click on Generate.
Make every control describe its real effect
“View details” can show your application's task record and sanitized error information. “Stop waiting” can close a local panel or defer monitoring. The current API does not expose a generation-cancellation endpoint, so a local action must not claim that upstream execution or billing has stopped.
Enable result retrieval after succeeded, not after a timer reaches an arbitrary threshold. Once the file is downloaded, inspect the media before labeling it ready for publishing. A clip thumbnail is not proof of complete playback; see the poster versus first-frame guide.
Exercise slow and interrupted states
Use fixtures for a task that remains processing, a throttled read, a lost connection, a failed task and an unknown outcome. Confirm that visible labels distinguish those cases and the next action remains appropriate. Test keyboard access and screen-reader labels as well as animation.
Check that completed items stop consuming regular polling traffic and that reopening a pending item uses its existing ID. The model hub can supply current model information, but it does not supply task progress percentages. Accept the UI when its messages remain accurate even when the task takes longer than expected, without an invented countdown or a misleading 100% state.
Related tasks
- What Should You Check Before Retrying a Failed Image or Video Job?
- Why Does a Video Card Show a Black Image Before Playback?
Sources and scope
- Imagild API console: Current request limits, retry handling, authentication and asynchronous task lifecycle.
- Imagild model hub: Current model options and media-generation point costs, separate from request-rate budgets.
- MDN ARIA progressbar role: Indeterminate progress should not expose a fabricated numeric value and needs an accessible label.
Product instructions were checked against the current Imagild API contract and the public model catalog. Code, budgets and acceptance checks are illustrative; this article does not claim a new paid generation test, an uptime measurement, a search-volume study or a ranking of AI citations.
Prepared by Imagild Editorial, the company editorial team responsible for these guides. AI assisted research and drafting. This is documentation-based guidance with proposed examples, not a claim of hands-on testing or an independent product review. Report a correction.
Cover: Original Imagild product screenshot captured on 28 October 2026; model availability and prices may change. This image is not an API response or a performance benchmark.