
HTTP 202 means Imagild accepted your generation request; it does not mean your image or video is ready to download. Save the returned data.id, then read that task until data.status reports an outcome. Show a finished preview only after successful generation and a valid authenticated media download.
The distinction matters when a request returns quickly. Your customer has started a paid operation, but the application still needs to discover what happened. The Imagild API console documents this asynchronous contract.
Key Takeaways
- A successful submission response acknowledges a task, rather than delivering a synchronous image array.
- Read task status, usage state and output availability as separate signals.
- Do not turn a queued task into a second generation merely because a preview is still empty.
Give each response a precise meaning
MDN explains HTTP 202 as acceptance before processing has completed, with no guarantee that processing will ultimately succeed. Your interface should therefore distinguish “accepted” from “completed.”
For Imagild, the submission route is POST /images/generations or POST /videos/generations, beneath https://imagild.com/api/relay/v1. A 202 response wraps the task record in data. The following small example illustrates selected fields, not a complete response or a real task to query:
{
"data": {
"id": "11111111-1111-4111-8111-111111111111",
"status": "queued",
"usage": {"points": 10, "state": "reserved"},
"output_url": null,
"poll_after_seconds": 5
}
}
Here, reserved does not mean the preview is available. It describes a billing reservation while work remains unresolved. Store the task ID against the same application operation that supplied your Idempotency-Key.
Use a state table instead of a single success flag
| Observation | What it establishes | Appropriate interface action |
|---|---|---|
| Submission returns 202 and a task ID | The operation was accepted | Save the ID and display accepted status |
| Task is initializing, submitting, queued or processing | No successful output is reported yet | Continue bounded status reads |
| Task reports succeeded | Generation has succeeded | Fetch authenticated content and verify it |
| Task reports failed | Generation failure is reported | Show the error and inspect usage state |
| Task reports unknown | Upstream outcome needs reconciliation | Preserve identity and request review |
Do not infer a percentage from the names in this table. The current Imagild response does not establish an exact progress percentage, queue position or completion-time estimate.
Other providers also distinguish running predictions from completed predictions; Replicate's lifecycle reference illustrates that general design. Its cancellation, deadline and billing rules are separate from Imagild's contract.
Turn acceptance into a visible workflow
For a product-card image, your own application can use three labels: “Request accepted,” “Generating,” and “Checking download.” These are interface labels you implement, not new API statuses.
When the record reaches succeeded, request GET /tasks/{id}/content using a valid Bearer key. Active membership remains required for this read and the download. Do not render the private output_url as if it were an anonymous public image URL.
Only after receiving media bytes should you mark the asset ready for your user. If the download instead contains HTML, use the download-content diagnostic; an extension such as .png does not verify the payload.
Check both technical completion and spending
Current image tiers in the model hub cost 10 points at 1K, 15 at 2K and 20 at 4K. A missing preview is not evidence that another attempt is free. Read data.usage.state alongside status: a confirmed failed generation releases its reservation, while an unknown outcome needs reconciliation.
Before shipping the interface, replay three saved, sanitized response examples: queued with no output, succeeded with content available, and failed with an error. Verify that only the successful case enables download and that none of the three automatically submits another generation.
The failed-job next-action guide helps frame the recovery message. The customer should understand whether the application is still waiting, needs an access correction, or has received a confirmed generation failure.
Related tasks
- What Should You Check Before Retrying a Failed Image or Video Job?
- Why Does My Downloaded Image Open as a Web Page?
Sources and scope
- Imagild API console and integration guide: Current Imagild entry point, active-member access, asynchronous task reading and authenticated output contract; implementation facts cross-checked in lib/relay and the 2026-10-28 claims manifest.
- Replicate Prediction Lifecycle: General distinction between an ongoing prediction and its reported outcome. Its provider deadlines, cancellation, billing and retention rules are not attributed to Imagild.
- MDN HTTP 202 Accepted: HTTP 202 acknowledges processing acceptance without establishing completed or ultimately successful work.
- Imagild model hub: Current image resolution point rates and H3 video point schedule; no uptime, comparative price rank or output quality guarantee is inferred.
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.