TTS stream responses are documented as raw audio streams
The Build docs now spell out that POST /v1/audio/stream returns chunked raw audio, not JSON, with clearer response content types and codec notes.
The Build docs now make the POST /v1/audio/stream response shape explicit: it returns chunked raw audio, not a JSON envelope.
That sounds obvious if you have already integrated the endpoint. It is not obvious if you are reading an API reference generator that tries to treat every response as structured JSON. The docs now call out the audio content types and the audio/L16 PCM behavior more directly.
Read the Build changelog entry for the exact clarification. No runtime behavior changed here. This was a docs and contract-accuracy fix so new integrations stop guessing at the stream body.