{"api_id":"pika/pika-audio/pika-soundtrack","vendor":"pika","function":"pika-soundtrack","category":"audio","name":"Pika Soundtrack","description":"Generate a synchronized soundtrack for a video. The visuals and duration are untouched; only the audio is replaced.","display_pricing":{"components":[{"role":"output","unit":{"type":"second","quantity":1,"included":0},"starting_at":{"spec":{},"sell_usd":"0.005","micro_usd":5000},"price_tiers":[{"spec":{},"sell_usd":"0.005","micro_usd":5000}]}]},"call":{"method":"POST","path":"/v1/media/pika/pika-audio/pika-soundtrack"},"input_schema":{"additionalProperties":false,"properties":{"video":{"description":"URL of the source video. H.264 or HEVC, at most 1200 seconds (20 minutes) and 2GB. The container does not matter and the URL needs no file extension — the codec is read from the file itself. Its visuals and duration are preserved — only the soundtrack is replaced. A source below the model's working resolution for its aspect ratio (about 576x320 for 16:9) is upscaled to it, so a very small video still produces audio, just with less for the model to work from.","media_kinds":["video"],"pattern":"^https?://","title":"Video","type":"string"},"instruction":{"anyOf":[{"maxLength":10000,"type":"string"},{"type":"null"}],"default":null,"description":"What the audio should be — sound effects, ambience, dialogue, music, or overall style — and how it should follow the visible action. Reaches the model verbatim. Omit it, or leave it blank, for natural sound design derived from the video alone.","title":"Instruction"},"seed":{"default":-1,"description":"Deterministic seed. -1 picks a random one.","maximum":2147483647,"minimum":-1,"title":"Seed","type":"integer"}},"required":["video"],"title":"VideoToAudioRequest","type":"object"}}