[KongchangAI]
Company

Burning Bridges Studio

A startup studio focused on AI-generated products, known for designing the asynchronous waiting experience of AI generation; its products support photo-to-video generation.

Timeline (last 90 days)

Oct 20

Burning Bridges Studio models AI generation work as a persisted job record (including fields such as id, status, createdAt, updatedAt, outputUrl and errorCode) rather than a pending HTTP request.

Unverified50%
Oct 20

For slow AI jobs, replacing long-lived requests with persisted job records lets a job still belong to the account and be recovered after the user closes the tab, while the UI can poll or subscribe to status.

Unverified50%
Oct 20

When a job fails, credits or quota should be automatically refunded, the job should remain visible with a clear reason rather than silently disappearing, and a retry should create a new job rather than modify the failed one.

Unverified50%
Oct 20

In the Burning Bridges Studio case, the generation result is a 15-second, 16:9 MP4 file that users can download after previewing.

Unverified50%
Oct 20

To prevent double charges caused by network interruptions, the client should have a brief recovery window after submitting a job (Burning Bridges Studio uses 20 minutes), during which it can find the existing job instead of resubmitting, and should clearly tell users not to resubmit.

Unverified50%

All Facts (5)

Source Articles