AssemblyAI Voice Agent API vs CueLight

AssemblyAI Voice Agent API

One $0.075 a minute rate covers speech-to-text, the managed model, speech output, recordings and hosting, and the socket has an AsyncAPI file, typed error codes and a 30-second resume window. No concurrent-session limit is published, the status page has no Voice Agent component, and phone calls are inbound only through a Twilio trunk the customer owns.

CueLight

CueLight gives a voice pipeline a local end-of-turn probability every 20 ms from audio alone, with a 1.9 MB model and Apache-2.0 weights. It is one day old, English only, installs from Git rather than PyPI, has no CI or changelog, and its accuracy figures are the author's own.

AssemblyAI Voice Agent API

1. Send session.end before closing the socket. Without it the session stays resumable, and billed, for 30 seconds
2. Retry only at_capacity, concurrency_exceeded and internal_error, with backoff. Every other session error code is fatal
3. Send tool.result inside the reply.done handler, not straight after tool.call
4. Mint a one-time token with GET /v1/token for browser clients and never ship the API key to a page
5. Deduplicate webhooks on event_id, and verify X-AAI-Signature against the raw body bytes because a 401 is not retried

CueLight

1. Feed 320 float32 samples at 16 kHz mono per push. Resample first, or the call raises an AssertionError
2. Spell the preset as snappy or careful. Any other name silently uses default thresholds
3. Call clear() before a new call or session. Pass reset_on_turn_start=False if your pipeline already resets per turn
4. Pin weights by downloading them to a local folder and passing that path to from_pretrained
5. Don't use it to identify speakers or clone voices. The training data's conditions forbid it