CueLight vs ElevenLabs Agents API + MCP

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.

ElevenLabs Agents API + MCP

API keys scoped to endpoint groups, with per-key credit limits and service accounts. $0.08 a minute excludes LLM tokens and carrier minutes.

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

ElevenLabs Agents API + MCP

1. Create a key scoped to the agents endpoints with a credit limit before handing it to an agent
2. Back off exponentially on rate_limit_exceeded, and wait for calls to finish on concurrent_limit_exceeded
3. Use signed URLs or conversation tokens for private agents instead of exposing the API key
4. Set platform_settings.privacy.retention_days where you don't need 2 years of history
5. Pick a low-latency LLM, the pipeline waits on it every turn