Developer Changelog

Introduces the ability to delete sent chat messages via chat.deleteMessage(), and several improvements to WebRTC connection recovery, WebCodecs fallback reliability, and audio device selection on mobile, plus expanded WebRTC video support for Firefox, Island Browser, and Chrome on Linux. Fixes issues with missing device-change events, stop events not firing for video and share state changes upon leaving or reconnecting, and a virtual background web worker that persisted after a session ended. Native Video SDK versions prior to 2.2.10 and Video SDK for web versions prior to 2.2.5 are not compatible with version 2.6.0 or later for command channel message exchange.

Breaking changes include a reduced raw audio send buffer requiring calls to stay under 100 ms, rejection of clients that don't support the new OnHold feature, and significant dependency updates. Removed a number of deprecated methods across classes such as ZoomVideoSDKDelegate, ZoomVideoSDKUser, and ZoomVideoSDKVideoHelper, along with several obsolete error codes in ZoomVideoSDKErrors. Additions include a full OnHold holding-state system in the join flow with helper classes, user management interfaces, and callbacks, support for specifying a camera device when starting video preview, and a new canSwitchAudioSource() method in ZoomVideoSDKAudioHelper. Deprecated several methods related to video preview and speaker management in favor of updated alternatives Microphone and speaker testing is now available both before joining a session and after audio has been stopped.

Bug fix for WebRTC session decode failure. Action required, upgrade to this version or later to fix issues.

Introduces an AI Agent Skill for development assistance that can be installed via npm to help with developing and debugging Video SDK applications. Enhanced command channel to support cross-subsession messaging between the main session and subsessions, as well as direct messaging capabilities. Added support for additional transcription languages. Fixes issues with virtual background enablement when CSP doesn’t include the unsafe-eval directive. Resolves screen share tab audio issues in certain edge cases.

Minor bug fixes.

Version incompatibility for command channel. Upgrade to the latest version of the SDK for the best experience. Support user failover status change. Add a pre-connection (two-step join) flow to speed up the actual session join. Joining or leaving a subsession now triggers the onSubSessionStatusChanged and onSessionLeave callbacks in useSdkEventListener.ts. Increase the command channel payload size limit to 1024.

Version incompatibility for command channel. Upgrade to the latest version of the SDK for the best experience. Breaking changes include updated audio configuration API naming for consistency, and modifications to external video source handling with new I420Limited and I420Full color range support. Adds a pre-connection flow to accelerate session joins, machine-answer detection for PSTN call-outs, and comprehensive audio QoS statistics. Additional improvements include user failover status tracking, expanded command channel payload limits (now 1024 bytes), and new callbacks for subsession status changes that trigger when users join or leave subsessions.

Adds RTMP ingestion for external streaming sources and a customAudioConstraints audio option; improves shared screen rendering, AMD environment support, and virtual background on low-performance devices; fixes screen sharing on Safari 27 / macOS 27, a Traditional Chinese language code, and a missing userKey field.

Added AI-friendly docs and Broadcast Streaming support. Fixed `enableMicOriginalInput`.

Add Android PSTN callout status visibility and support for voice translator.

Fixed bugs with stopShareView and stopBroadcast methods.

Rollback of versions 2.5.5 and 2.5.6, with 2.5.7 as the fix.

Adds real-time voice translation, allowing a speaker’s voice to be translated into other languages during a session. Adds greeting audio language support for stream.inviteByPhone() to customize PSTN dial-out greetings. Enables 1080p video resolution in WebRTC mode. Automatic CDN fallback for asset loading, better session join stability, a new dialType property to identify PSTN dial-in or dial-out users, and enhanced audio reliability under poor network conditions. Fixes error with stream.stopAnnotation() on the viewer side and a blank screen issue in stream.startShareView() caused by incomplete state cleanup.

Adds enhanced QoS statistics for Web SDK consistency, auto-loading multi-video stream support, multi-language telephony "welcome" messages, and new AI-friendly API documentation. Breaking changes to annotation and cleanup workflows.

Addresses a WebRTC issue that caused session negotiation failures and fallback to WASM on Chrome 147 and later. If you rely on WebRTC video, we strongly advise that you upgrade to ensure compatibility with upcoming Chrome versions.

Adds support for subsessions with new interfaces for managing session types, user lists, and session status. Introduces audio control features including muteAll and unmuteAll. Fixes issue where ZoomVideoSdkUser was returning incorrect user information.

Fixed an issue where the iOS onUserLeave callback returns data inconsistent with Android.

Adds comprehensive whiteboard support, including new listener callbacks for tracking whiteboard share status changes and export events, a WhiteboardStatus enum, user-level whiteboard status checking, and a complete WhiteboardHelper class with methods to start, stop, and export whiteboards. Also introduces a WhiteboardView component and an audio session configuration interface. The update removes x86_64 iOS simulator support (now requiring arm64) and fixes several iOS issues including VideoView memory leaks, subscription problems, pre-session video rotation errors, and picture-in-picture functionality.

Adds support for appearance touch-up and limiting the maximum resolution of shared media. Deprecates unused error enumerations. Changed exportLog to export client logs.

Fixes errors to improve virtual background CPU optimization, hardware capability detection, video detachment stability, and add proper error handling for audio operations during session transitions.