remove words
Text-based cutting: remove words by transcript index or by text match, mapped to frames, gaps closed. After a cut the indices shift, so get_transcript is re-read before the next call. The pause around a removed word goes with it, leaving one beat. Free.
Authorizations
An API key (es_live_…) from Settings, under API keys, or an OAuth access token.
Headers
Retry a paid dispatch safely: the same key returns the job already dispatched instead of charging again.
The workspace this one call acts in, when it is not the default: for a key or a connected agent, one of the workspaces the user allowed it.
Body
The timeline id.
Word indices from get_transcript; a [start, end] pair cuts that whole span.
Cut every occurrence of these single words (case and punctuation insensitive). For a phrase, pass its word indices from get_transcript as a [start, end] span.
The beat left where a word was: tight about 0.2 s, balanced about 0.3 s (default), loose about 0.4 s. A word that stays is not cut into.
tight, balanced, loose The speech the indices came from: the same source get_transcript was read with (all included), or the indices land on the wrong stream.
voiceover, dialogue, music, auto, all The revision get_timeline returned. A stale one fails with revision_conflict instead of applying; so does no revision once a user edited the timeline in the app since this connection last read it.
Optional: one sentence shown to the user beside this action.
600
