Skip to main content

Compatibility

Platformv0.9.11Recommended SDKv0.1.3Supported SDK range>=0.1.0 <0.2.0

Platform

The current SDK release is tested against the platform release shown above. SDK versions in the supported range are expected to interoperate with this platform release.

Platform versionRecommended SDKSupported SDK range
0.9.110.1.3>=0.1.0 <0.2.0

Platform release notes always include a compatibility section. See /release-notes.

Browsers

CapabilityRequired browser feature
text, audio_playback, listen, browser_actionsWebSocket, fetch, modern Promise / async
voice_conversationAll of the above plus RTCPeerConnection

Every evergreen browser (Chrome, Edge, Firefox, Safari) released in the last two years meets the requirements. IE and legacy Edge are not supported.

getUserMedia({ audio: true }) is required for voice_conversation and is gated behind a user gesture in some browsers; place startCall inside a click handler.

React

react versionStatus
>=18Supported (uses useSyncExternalStore).
<18Not supported.

TypeScript

>=5.0 is recommended. The SDK ships full .d.ts (ESM) and .d.cts (CJS) typings.

Node

The SDK is browser-only at runtime. Node is supported for builds and tests via engines.node >= 18.

Backend requirements

Your backend must include these fields in the session-token response so the SDK can complete a voice call:

FieldPurpose
sessionIdServer-issued session identifier.
wsTokenToken used to authenticate the channel WebSocket.
wsUrlURL of the channel WebSocket.
telproDomainTelPro routing hint; also seeds default ICE servers when iceServers is omitted on DelphiClient.
webrtcGatewayUrlWebSocket URL of the WebRTC gateway. Required for voice_conversation.

The SDK does not expose a webrtcGatewayUrl config option — the value is per-session and comes from the token response. See Configuration.