Before rendering a video you need to generate visuals for your slides from the generated text.

🚧

Fix with an LLM

Set fixWithLlm to true and an LLM pass runs first: missing text is written, over-long speech
is shortened, and content is fitted to the slide layout. It needs LLM access on the account —
without it, or once the allowance is spent, the call answers 429 and no slides are generated.

This flag was previously documented as fixWithGpt, a name the handler never read.

curl -X POST 'https://apis.elai.io/api/v1/story/generate-slides/66a1f0c2b7cd13032a9c6002' \
  -H 'Authorization: Bearer YOUR_API_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{ "fixWithLlm": true }'