Voices only: isolate each take's voices (costs credits)

Seedance clips carry their own music and effects almost all the time, so the film's score ducked under them stays down for the whole film. Voices only sends each take's soundtrack to ElevenLabs audio isolation and keeps the voices as the take's voice track (voicePath, with speech: where it speaks, in clip seconds). The player then plays the voices at the voices level and the clip's full sound under them at filmTexture (the levels: mix on GET /v1/storyboards/{id}, set with the mutation {op: "setMix", patch: {film, filmTexture, voices, music, ambience, effects, voiceOver}}, each 0-1.5, null back to the default), and the music, place beds and effects dip about 10 dB under speech only (the lines and the voice tracks' spans), never under the clips' own sound. {op: "setVoicesOnly", target, takeId, on: false} plays a take's full sound again, free (the track is kept). Without take_ids: every take in use (segments and ending shots; a waiting loop keeps its own sound) that has no voice track yet. Costs credits: audio_isolate_10s per 10 s of audio (see /v1/pricing). Quote first with dry_run: true: credits is the most it costs, per take in takes; available: false comes with a reason (nothing to do, a run going, or the price not set yet). Without dry_run it answers 202 { task_id }: poll GET /v1/tasks/{id}. Each take is charged after its track is on the storyboard, never more than quoted, through your token so its daily cap applies. Paid tier only (402 audio_requires_purchase on trial credits). A storyboard whose owner approves agent spending answers 403 except for the quote.