Programmable WebRTC rooms, predictable latency

Embed live multi-party audio and video in your app with Synton's Selective Forwarding Unit (SFU) engine, 3-layer simulcast, and room-scoped Bearer tokens.

room-simulcast.js · livekit-client v2.22.3
const room = new LivekitClient.Room({
  adaptiveStream: true,
  dynacast: true,
  publishDefaults: { simulcast: true }
});
await room.connect("wss://rtc.synton.cc", "<TENANT_TOKEN>");
await room.localParticipant.setCameraEnabled(true);
Simulcast Rung Resolution Target Bitrate Dynacast State
Layer HIGH 1280 × 720 (30 fps) 1,500 kbps Active for pinned speaker
Layer MEDIUM 640 × 360 (24 fps) 500 kbps Active for visible grid tiles
Layer LOW 320 × 180 (15 fps) 150 kbps Paused when tile hidden
Illustrative 20-participant standup profile using Synton's client SDK (sfu-client.js) and WSS signaling gateway (/rtc).
Plane Health All planes nominal Live readiness probe
Simulcast Rungs 180p / 360p / 720p Dynacast pauses unviewed layers
Signaling & Auth Room Bearer JWT Issued by app backend for /rtc
Media Codecs Opus + VP9 / H.264 WASM decoders available

Capabilities Organized by Task

01 / SIMULCAST

Multi-Party Video Collaboration

Support 20-participant engineering standups and interactive panels without overwhelming subscriber downlinks. Publishers send 180p, 360p, and 720p layers while Adaptive Stream selects only the resolution matching each rendered video viewport.

02 / DYNACAST

Automatic Bandwidth Control

When every subscriber pauses, minimizes, or scrolls a participant off screen, the SFU signals the publisher to pause higher simulcast encoders until a viewer subscribes again.

03 / NAT RELAY

Restricted-Firewall Enterprise Calls

Corporate callers behind symmetric NAT or strict UDP egress fall back transparently to regional TCP/TURN relay endpoints on the same media plane while preserving audio intelligibility.

04 / ISOLATION

Tenant Pipelines & Recording

Dedicated WebRTC media forwarding pipelines, per-project API credentials, and isolated recording egress are provisioned per workspace under enterprise agreements.

Three-Step Integration Workflow

STEP 01

Mint a Room Token

Your application backend validates user permissions and signs a short-lived Bearer token scoped to a single room identifier and participant identity.

STEP 02

Connect via Client SDK

Load sfu-client.js, connect to wss://rtc.synton.cc, and publish microphone and camera tracks with simulcast enabled.

STEP 03

Adapt & Verify

Adaptive Stream scales track quality per layout tile automatically. Inspect runtime codecs and regional corridors in Guides and Network.

Engineering Questions

  • How do Adaptive Stream and Dynacast keep crowded rooms fast?

    Each subscriber requests only the simulcast layer matching its visible tile size; unviewed video rungs are paused upstream so downlink and CPU stay bounded. Read the full explanation.

  • What happens when a corporate firewall drops UDP media?

    ICE negotiation shifts smoothly to regional TCP/TURN relay endpoints and prioritizes voice packets over high-bitrate video layers. See firewall fallback details.

  • Why must room signaling tickets expire quickly?

    Short-lived tokens minted by your backend right before joining prevent replay outside the intended room window and revoke cleanly on HTTP 401/403. Inspect API auth rules.