NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
pub.dev, #3032 most downloaded on pub.dev
ZegoExpressEngine Flutter SDK is a flutter plugin wrapper based on ZEGO Express RTC SDK (Native & Web), providing live video and real-time audio/video services.
Latest stable
3.25.0
Last release 14 days ago
06 Nov 2026
Ships fairly regularly
a new release about every 2 weeks
Most releases are documented
notes for 43 of the last 60 stable releases
6 versions withdrawn
withdrawn after publishing
7 years old
357 releases · first in 2020
Drag across the ruler to compare, or tap a release to open it.
Added AI Echo Cancellation (AEC) Lightweight Mode - All Platforms
Added AI Echo Cancellation (AEC) Lightweight Mode - All Platforms
setAECModeSupports Video Noise Reduction Function - iOS/Android/Flutter
setVideoDenoiseParamsSupports Returning Face Recognition Status from Camera Capture - iOS/Android
Supports Adding Mixed Stream Output to Target Room - All Platforms
startMixerTask targetRoom paramOptimized Noise Reduction Effect in Low Light Enhancement Function - iOS/Android/Mac/Windows
Fixed Issue Where Media Player Fails to Load URL Links with Spaces
Fixed Issue Where iPad Cannot Connect to Bluetooth Headphones
Fixed Issue Where MacOS Throws State Error After Disconnecting the Current Media Device
Fixed Issue Where Media Player Crashes Occasionally When Local Cache Is Enabled
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Fix the issue of crashing when custom video processing use SurfaceTexture video type.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Support for Linux Platform in Flutter Framework
Support for Linux Platform in Flutter Framework
Support for Alpha Data Layout on Top-Right of RGB Data in Media Player
When playing transparent effects, the media player now supports an enum ZegoAlphaLayoutType > RightTop, allowing the alpha data to be placed on the top-right of the RGB data. When using this enum, only a 0.5x scale ratio is supported.
Relevant APIs: loadResourceWithConfig, ZegoStreamResourceMode > Custom, ZegoAlphaLayoutType > RightTop.
Support for Custom Stream Resource Types Before and After Audience Join/Leave
Allows setting different stream resource types before and after an audience member joins or leaves the stage, providing more flexible stream handling. You can choose from RTC streaming, ultra-low latency live streaming (L3), or CDN streaming.
For example, in live streaming with audience interaction, the audience can use L3 streaming before joining the stage, switch to RTC streaming when interacting, and revert to L3 streaming after leaving the stage.
Relevant APIs: startPlayingStream, ZegoPlayerConfig > customResourceConfig.
Removal of Default Screen Sharing Permissions in AndroidManifest.xml
Note: After version 3.17.0, if you implement screen sharing functionality without declaring the required permissions, you must now manually declare them to avoid compatibility issues. For any compatibility concerns in live applications, please contact ZEGO technical support.
To align with privacy regulations, the default screen sharing permission has been removed from the AndroidManifest.xml. Developers must manually declare permissions:
FOREGROUND_SERVICE permission.FOREGROUND_SERVICE and FOREGROUND_SERVICE_MEDIA_PROJECTION permissions.Refer to the Screen Sharing documentation for more details.
userName Field in loginRoom Now Optional
When calling the loginRoom API to enter a room, the userName field is now optional, previously it was a required field.
Relevant APIs: loginRoom.
Improved Stereo Voice Changing Effects and Enhanced Music Pitch Quality
setVoiceChangerParam.Fixed Issue with Parsing H.265 Streams Containing Abnormal Cropping Regions
Resolved a Crash in the Monitoring Module During Deinitialization, Which Caused the Main Thread to Hang
Fixed a Crash That Might Occur When Starting Scoring for Copyrighted Music
Adjusted Lifecycle and Fixed Known Crash Issues
Fixed Known Issues in the Media Player
Fixed the problem that the camera cannot be opened in certain Android scenes
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Deprecated the CDN Plus configuration in the stream pulling interface
Note:
ZegoVideoSourceTypeSecondaryCamera is introduced to identify the video from the second camera. It supports setting both front and rear cameras as video sources and streaming them separately using the [setVideoSource] interface. Developers can use [useFrontCamera] to switch the front and rear views corresponding to ZegoVideoSourceTypeCamera and ZegoVideoSourceTypeSecondaryCamera. This capability can be applied to scenarios such as dual-camera live streaming. For more information, refer to Multi-source Capture.
Relevant APIs: setVideoSource, useFrontCamera, ZegoVideoSourceType > ZegoVideoSourceTypeSecondaryCamera, ZegoVideoSourceType > ZegoVideoSourceTypeCamera.Note:
Note:
Note:
Note:
Note:
dlopen intercepting libnativewindow on some phones.Nothing published for this version
Nothing published for this version
Nothing published for this version
### Bug Fixes 1. Fixed MediaPlayer error.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
…interface is deprecated. Please use [ZegoMediaPlayer.enableVoiceChanger] instead.
Note: Please contact ZEGOCLOUD technical support if you need to use this feature.
Sound data that supports in-ear monitoring can be played from the speakers.
The media player has added the [enableVoiceChanger] interface, which supports enabling voice changer effects for the audio output of the media player, and allows selection of the desired pitch-shifting effects.
For related API, please refer to ZegoMediaPlayer > enableVoiceChanger
Optimize the noise reduction effect of "Balance Mode" in AI scene-based noise reduction, further enhancing the clarity and stability of the human voice without compromising performance, and achieving cleaner noise suppression.
Optimize video hardware encoding and decoding, increase the range of compatible devices, and enhance the performance of encoding and decoding at high resolutions and high frame rates.
Fixed the issue of probabilistic crashes when stopping the audio and video engine with the input source from the auxiliary stream in the case of using the main stream for pushing.
Fixed the occasional playback failure issue when playing network materials with the media player.
Fixed the issue of probabilistic crashes during hardware encoding initialization.
Fixed the issue of playback problems when hardware decoding CDN live video streams containing B-frames.
Fixed the issue of accessing a wild pointer during camera enumeration.
Web removes use of innerHTML interface to reduce security risks.
To enhance the playback experience during voice changing, the media player [ZegoMediaPlayer.setVoiceChangerParam] interface is deprecated. Please use [ZegoMediaPlayer.enableVoiceChanger] instead.
For related API, please refer to ZegoMediaPlayer > enableVoiceChanger
Nothing published for this version
Nothing published for this version
Note: Starting with this release, iOS 11.0 and earlier versions are no longer supported.
Note: Starting with this release, iOS 11.0 and earlier versions are no longer supported.
Starting from 2024-04-29, all apps on the App Store must support iOS 17.0 version. For details, please refer to Apple Developer Website Official Instructions.
Note: Please contact ZEGOCLOUD technical support if you need to use this feature.
ZegoVoiceChangerPreset added enumeration values for two voice-changing effects, Autobot and OutOfPower, to enrich the voice-changing effects.
For related API, please refer to [setVoiceChangerPreset]
For related API, please refer to [enableViewMirror]
ZegoMixerTask added new parameter mixImageCheckMode, used to control whether backgroundImageURL, inputList.imageInfo.url, or watermark.imageURL image resources fail to verify whether the mixing task can be initiated normally.
This function is not enabled by default (the default value of mixImageCheckMode is 0), which means that image verification is strictly performed, that is, the original restriction rules of parameters must be met before the mixed flow task can be initiated normally.
For related API, please refer to [startMixerTask]
The AI voice changing function has certain requirements on the performance of the running device. You can use the [isAIVoiceChangerSupported] interface to determine in advance whether the device can support the AI voice changing function.
For related API, please refer to [isAIVoiceChangerSupported]
Optimize the super-resolution effect and reduce the sharpness of the algorithm, thereby improving the subjective quality when the original picture has noise, there are faces in the picture, etc. For example, the flaws on the anchor’s face will not be highlighted, and the anchor’s hairline will not become more obvious.
Optimize the frame rate performance of the super-resolution algorithm on Qualcomm chips.
Fixed an issue where the release timing of undefined global variables caused the process to exit abnormally.
Fixed the issue where some models use AudioTrack to play audio, resulting in abnormal sound.
Fixed the problem of using the AI voice changing function to occasionally cause the audio and video to be out of sync, and causing words to be swallowed instantly when starting and stopping.
Fixed the issue where the iPad collects audio through the microphone and occasionally has no sound.
Fixed the problem of multi-thread concurrency when destroying a camera, causing a crash when accessing a null pointer.
Fixed the problem that the mediaPlayer takes a long time to open files in M3U8 format.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Fixed the problem of using TextureView to render video frame on the iOS platform to set viewMode abnormally.
Nothing published for this version
Nothing published for this version
Fixed the problem of data loss when reporting quality or billing failed.
Nothing published for this version
Fixed an issue where publishing a screen share stream on the web platform failed.
This feature is only available for use during internal video capture.
Note:
This feature is only available for use during internal video capture.
When this feature is enabled, there may be delays or cropping of the image, so please use it accordingly.
The new video stabilization feature is added to reduce the impact of camera shake during internal video capture and improve the quality of video capture.
For related API, please refer to [setCameraStabilizationMode]
Note:
The security of this feature is slightly lower compared to traditional methods. Please use it with caution.
When using this feature, set ZegoCDNConfig.protocol to quic.
[ZegoCDNConfig] adds the [quicConnectMode] attribute, which allows developers to use the QUIC protocol for CDN streaming. Set quicConnectMode to 1 for QUIC connection mode, enabling 0-RTT connection and fast service activation. Currently compatible with CDN live streaming products from Huawei, Wangsu, Tencent, and other vendors.
This feature is not enabled by default (quicConnectMode is set to 0, indicating normal connection establishment).
For related API, please refer to ZegoCDNConfig > quicConnectMode
Note: This feature only takes effect when initiating a retweet. If there is a disconnection during the retweet process, the SDK will maintain the retry logic and there will be no callback notification in this case.
When initiating a retweet task, you can set the timeout for the retweet CDN through the [addPublishCdnUrl] interface to monitor if the stream exists. For example, if the developer has initiated a retweet task but the stream has not started streaming yet, after the timeout set, the SDK will return a callback notification indicating that the stream does not exist.
This callback notification will only be sent to the retweet initiator, not the streaming initiator. If the retweet initiator and the streaming initiator are not the same user, it is recommended for developers to initiate the retweet from the server side and receive this notification.
For related API, please refer to addPublishCdnUrl
[ZegoDataRecordProgress] adds the [quality] attribute, which can be used to callback the quality data of the recorded file, such as frame rate and bit rate, during the local recording process.
For related API, please refer to onCapturedDataRecordProgressUpdate
Note: The external capture function and pre-processing function cannot be used at the same time, otherwise abnormal images may occur when playing streams.
After enabling the external capture function, you can use the [setLowlightEnhancement] and [enableColorEnhancement] interfaces to separately enable low-light enhancement and color enhancement to adjust the captured images according to your business needs.
For related API, please refer to setLowlightEnhancement, enableColorEnhancement
Note: Please contact ZEGOCLOUD technical support if you need to use this feature.
When some users in the room do not support the H.265 format, the streaming end that supports it will fall back to the H.264 format and republish stream.
Optimize the callback notification logic of the media streaming engine, add error callbacks for unsupported audio sampling rate (for example, not supporting a sampling rate of 24K), and help developers quickly locate problems.
For related API, please refer to onMediaDataPublisherFileClose
Optimized color enhancement algorithm performs better than previous versions in scenes with high color saturation.
Note: Please contact ZEGOCLOUD technical support if you need to use this feature.
Fix the problem of abnormal time consumption of interface calls caused by not deinitializing the SDK after long-term use.
Fixed compatibility issues with hardware encoding and decoding, addressing occasional crashes.
Fix known compatibility issues and null pointer problems.
Fixing an issue where the engine occasionally initializes with incorrect states.
Nothing published for this version
Fixed the issue where the video rendering black screen on Mac10.15 safari 13.0.2.
Fixed the issue where the UI would occasionally freeze if a network abnormality occurs when initializing the SDK.
Fixed the issue where the UI would occasionally freeze if a network abnormality occurs when initializing the SDK.
Fixed the issue where UI freezes may occur in a very low probability if the network is abnormal when switching networks.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Please contact ZEGOCLOUD technical support if you need to use this feature.
Note:
Please contact ZEGOCLOUD technical support if you need to use this feature.
The plugin cannot be used alone and must be used with Express SDK.
Support for copyright-music function pluginization, when the developer's business scenario only needs to update the copyright-music related, you can independently integrate the plugin without updating the Express SDK, which can smoothly migrate.
Note: The function retrieves a real-time stream list inside the room. If the room service is disconnected, the results obtained may not be accurate.
Developers are supported to obtain the stream list inside the room from the client, which can be used to handle related business logic.
For related API, please refer to getRoomStreamList
Note: Please contact ZEGOCLOUD technical support if you need to use this feature.
Support is provided for adding silent frames to the audio and video streams that are pushed to the CDN. This can be used to avoid issues such as stuttering or audio-video synchronization problems caused by timestamp discrepancies.
Support for obtaining frame rate statistical information of the currently playing media file, which can be used for data display, anomaly monitoring, etc..
For related API, please refer to getPlaybackStatistics
Support local caching of network resources, so that if the same network resource needs to be played, cached data will be prioritized, enhancing user experience.
For related API, please refer to enableLocalCache, onMediaPlayerLocalCache
Fix the issue where certain models would display stripes on the screen when motion is occurring after enabling the super resolution feature.
Fix the issue of memory leak when destroying GPU resources in external filter blending mode for some phones.
Fix the issue of stuck when playing and rendering streams on some devices.
Fix the issue of crashing when playing audio or video with a sample rate higher than 48K on certain Android devices.
Fix the issue of occasional crashes when calling the [enableAudioCaptureDevice] interface.
Nothing published for this version
Nothing published for this version
Note: If you need to use this feature, please contact ZEGOCLOUD business personnel.
Note: If you need to use this feature, please contact ZEGOCLOUD business personnel.
By applying leading coding and decoding algorithms and other video pre-processing capabilities in the cloud transcoding service, we continuously optimize the smoothness and clarity of video playback, significantly improving the image quality. This feature is suitable for the following scenarios:
Showroom live streaming scenes with high viewership. It ensures stable video transmission and high quality while saving bandwidth costs; without affecting the image quality, it can reduce the bitrate by about 30%.
Danmaku game live streaming, sports live streaming, and other scenes with rich color and texture details in the video content. Under the same bitrate conditions, it can provide a higher definition viewing experience.
For related API, please refer to ZegoMixerOutputVideoConfig > enableLowBitrateHD
For various cameras and other devices that capture images, if the colors appear grayish or have low saturation, we support enhancing the colors while preserving the natural skin tones. This will make the images more vibrant and brighter, creating a more realistic visual experience for the human eye.
For related API, please refer to enableColorEnhancement
Support sending real-time room messages to specified clients or client servers; message types are divided into normal and ordered, with the latter ensuring that messages are received strictly in order. This feature is suitable for scenarios where the anchor needs to manage the microphone positions in the room, for example:
Send messages to users who need to mute through the anchor client, and the receiving client will mute accordingly.
When the anchor wants to kick a user out of the room, send a message to the client server of the other party through the anchor client, and kick out the user.
For related API, please refer to sendTransparentMessage
Note: This feature only supports pre-processing of screenshots and does not support other processing such as rotation or watermarking.
When the video format output by the capture device is MJPEG, hardware decoding acceleration is enabled by default to prevent issues such as insufficient frame rate due to insufficient device performance.
This feature is suitable for use on capture devices with a 4K resolution mainly.
Note:
This feature is not enabled by default, meaning the server uses the default configuration values.
This feature may increase latency, so use it judiciously.
The automatic stream mixing interface supports setting a watermark to control the lower limit of the range for adaptive adjustment of the mixing server's stream cache. This helps maintain a balance between mixing time and video stuttering caused by unstable streaming from the source. This feature only takes effect on new input streams and does not affect input streams that have already started mixing.
For example, in a real-time karaoke KTV scenario, slight fluctuations in the streaming network from the source may cause mixing stuttering, which in turn increases the likelihood of stuttering for viewers. By adjusting the lower limit of the watermark, you can optimize the viewer's experience with stuttering, but this will increase latency.
For related API, please refer to ZegoAutoMixerTask > minPlayStreamBufferLength
Newly added support for using live streams as input streams for mixing; the URL of the live input stream supports both RTMP and HTTP-FLV protocols. This feature is suitable for mixing the RTC video streams of hosts' interactive broadcasting with cloud sports live streams, game live streams, etc., to achieve scenarios such as game or sports commentary in live broadcasting.
When using custom audio and video capture function and the corresponding audio capture sources have inconsistent delays, you can customize the audio offset value during mixing to achieve audio-video synchronization after mixing output, ensuring a better experience for the audience.
For related API, please refer to ZegoMixerInput > advancedConfig
The media player supports throwing relevant callback notifications to developers when the video resolution changes. This feature is suitable for scenarios where the resolution of the streaming screen changes multiple times and requires adjusting the encoding resolution on the streaming end and matching the rendering view size on the receiving end.
For related API, please refer to onMediaPlayerVideoSizeChanged
The sound effect player supports setting the streaming volume and local playback volume separately, ensuring that the volume on both ends, local and remote, is within an appropriate range.
For related API, please refer to ZegoAudioEffectPlayer > setPublishVolume, ZegoAudioEffectPlayer > setPlayVolume, ZegoAudioEffectPlayer > setPublishVolumeAll, ZegoAudioEffectPlayer > setPlayVolumeAll
Optimize server-side mix streaming and single-stream transcoding capabilities to improve encoding efficiency and achieve a 5% or more increase in subjective and objective video quality at the same bitrate.
After the user successfully logs in on device A, device A loses network connection. Then, the user logs in successfully on device B using the same userID. If the network connection on device A is restored and a reconnection is attempted, it will fail and throw error code 1002086, indicating that the userID is already logged in another device.
Nothing published for this version
Fixed the problem of false positives in the mobile terminal sleep detection module, affecting room re-login and push-pull stream retry logic.
Nothing published for this version
Nothing published for this version
Fixed the issue of a black screen appearing after turning on low-light enhancement.
Nothing published for this version
One column per quarter.