Rewrites the existing speech on the named slides following enhancement as instructions.
Replaces the retired POST /gpt/{videoId} and POST /gpt/{videoId}/{slideId}. Where those took one of ten fixed enhancement values, this takes plain-language instructions.
| Time | Status | User Agent | |
|---|---|---|---|
Retrieving recent requests… | |||
Rewrites the existing speech on the named slides following enhancement as instructions. To write
speech for a slide that has none, use
Generate speech for slides instead.
Replaces the retired POST /gpt/{videoId} and POST /gpt/{videoId}/{slideId}.
Migrating from the /gpt enum
/gpt enumThe old endpoints accepted one of ten fixed enhancement values. There is no enum any more —
phrase it as instruction text:
| Old value | Instruction text |
|---|---|
engaging | "Make this more engaging." |
professional | "Make the tone more professional." |
happier | "Make the tone warmer and more upbeat." |
shorter / longer | "Cut this by about a third." / "Expand this with more detail." |
rewrite | "Rewrite this from scratch, same meaning." |
The four structural values have no single-call replacement here. create-by-outline is now
Build a video from an outline; add-missed-texts, strecth-slides and
fit-speech reshaped the whole deck and have no direct equivalent — contact
[email protected] if you relied on one.
Notes
slideIds must be non-empty, and every id must exist on the video or the call returns 404
naming the missing ids. Speech is persisted server-side unless you set fromEditor — leave it
unset for API use.
Requires LLM access on the account — 409 if unavailable.