Processing and limits
Generation jobs are queued server-side and processed in order, with Growth plan jobs queued ahead of standard ones. A single variation takes roughly a minute; a batch is processed sequentially, so a batch of twelve takes several minutes rather than twelve separate waits.
Last updated
What affects generation time
- Batch size, since jobs are queued rather than run in parallel.
- Target duration, because a longer ad means more clips to conform and more audio to synthesize.
- Script retries, when the first read overruns the target and the script is rewritten.
- Queue position, which is where priority processing on the Growth plan helps.
- AI video generation, which is substantially slower than assembling existing clips.
Rate limits
Endpoints that call paid external services are rate limited per account, and authentication endpoints are limited more tightly. Normal use never reaches these limits; scripted hammering of the generation API will.
Frequently asked questions
Why is my batch taking longer than one variation?
Batches are queued and processed in order rather than in parallel, so the total time scales with the batch size. It is still far less than the sum of separate requests.
What does priority processing do?
It places Growth plan jobs ahead of standard jobs in the queue, which matters most when the system is busy or you submit a large batch.
Does closing the browser cancel generation?
No. Generation runs server-side and the results appear in the campaign when finished.