Changelog

All notable changes are documented here. RTCstack follows Semantic Versioning.

v0.1.0 — 2026-04-21

Initial release.

Docker Compose stack: LiveKit SFU, LiveKit Egress, coturn TURN, Caddy TLS proxy, Redis, MinIO
All services use pinned image versions with Docker healthchecks
docker/.env.example with all required variables documented

POST /v1/token — sign LiveKit JWTs with role-based grants (host, moderator, participant, viewer)
POST /v1/token/refresh — refresh near-expiry tokens
GET/POST/DELETE /v1/rooms — room lifecycle management
GET/DELETE /v1/rooms/:roomId/participants/:id — participant list, kick
POST /v1/rooms/:roomId/participants/:id/mute — mute audio or video
POST/GET /v1/rooms/:roomId/recording/start|stop — composite MP4 recording via LiveKit Egress
POST/GET/PUT/DELETE /v1/webhooks — webhook config CRUD with HMAC-signed delivery and retry backoff
POST/GET/DELETE /v1/rooms/:roomId/ingress — RTMP/WHIP ingress via LiveKit IngressClient
GET/POST /v1/rooms/:roomId/transcription/* — live + post-call transcription (requires STT profile)
HMAC-SHA256 request signing with 5-minute replay protection window
Rate limiting via @fastify/rate-limit

createCall({ token, url, tokenRefresher? }) — factory returning a Call instance
Call class wrapping livekit-client: connect, disconnect, toggleMic, toggleCamera, startScreenShare, stopScreenShare, switchDevice, sendMessage, sendReaction, setLayout, pin
Typed EventEmitter with full CallEventMap
MockCall + createMockCall() for testing without a LiveKit server
Bundle ≤ 80 kB gzipped

@rtcstack/ui-react — CallProvider, 7 hooks, 5 components (VideoConference, VideoGrid, ControlBar, ChatPanel, ParticipantVideo), CSS design tokens, dark/light themes
@rtcstack/ui-vue — CALL_KEY injection, 7 composables, 5 SFC components
@rtcstack/ui-vanilla — mountVideoConference() imperative API