Sign inSign up

n1snt/livekit-rs-voice

By n1snt

•Updated 30 days ago

A minimal, voice only LiveKit server in Rust.

Image
Languages & frameworks
Machine learning & AI
Web servers
0

1.1K

n1snt/livekit-rs-voice repository overview

⁠livekit-rs-voice

A minimal, voice-only LiveKit server in Rust — a drop-in replacement for the audio path of livekit-server. It speaks the same LiveKit wire protocol, so existing LiveKit clients, livekit-agents, and the livekit/sip + livekit/egress containers connect and work unchanged.

⁠How to use this image

docker run --rm -p 7880:7880 -p 7881:7881/udp -p 7881:7881/tcp -p 7882:7882 \
  -v "$PWD/livekit.yaml:/etc/livekit/livekit.yaml:ro" \
  n1snt/livekit-rs-voice:latest --config /etc/livekit/livekit.yaml

Minimal livekit.yaml:

port: 7880
keys:
  devkey: your-api-secret

For production, generate a key with at least 32 characters. See the full LiveKit configuration reference⁠.

⁠What's inside

  • LiveKit-compatible signaling: /rtc, /rtc/v1 WebSockets and the Twirp API
  • Voice-only media plane (audio forwarding, active-speaker detection)
  • SIP interoperability with the livekit/sip container (inbound + outbound) over Redis (psrpc)
  • Egress support via the livekit/egress container
  • Drop-in livekit-server Prometheus metrics

⁠Tags

  • latest — latest release
  • 1.13.5 — versioned release

Multi-arch: linux/amd64, linux/arm64.

⁠Documentation

⁠License

Apache-2.0. See LICENSE⁠ and NOTICE⁠.

Tag summary

Content type

Image

Digest

sha256:10e0ffbd2…

Size

16.6 MB

Last updated

30 days ago

docker pull n1snt/livekit-rs-voice