Xiaozhi Android Client

一个基于小智、xiaozhi-server的Android、IOS语音对话应用,支持实时语音交互和文字对话。现在是flutter版本,打通IOS、Android端。请同志们动动小手,点点小星星,予以鼓励。

Last verified:

Visit Xiaozhi Android Client

What is Xiaozhi Android Client?

xiaozhi-android-client is a cross-platform Flutter-based AI voice assistant client for Android and iOS that enables real-time voice and text conversation with AI assistants. Built using the Flutter framework, it supports deployment on iOS, Android, Web, Windows, macOS, and Linux with a single codebase, providing seamless voice对话 functionality through WebSocket connections to Xiaozhi AI server, Dify, and OpenAI services.

Key features include real-time voice calling with continuous conversation mode, text messaging, image message support, manual call interruption, push-to-talk functionality, AEC+NS echo cancellation for Android, live2d lip-sync animation, mood interaction features, multi-agent support, Qwen3 model thinking mode toggle, and automatic device registration via OTA. The app offers adaptive UI layouts, smooth animations, light skeuomorphic design, and supports multiple AI service configurations with the ability to add multiple assistants to the chat list.

The client is designed for users who want hands-free AI interaction, including personal assistance, educational tutoring, companionship, and multilingual communication support. It supports five languages (Mandarin, Cantonese, English, Japanese, Korean) for voice recognition and provides real-time speech-to-text conversion for enhanced accessibility.

As an open-source project under Apache-2.0 license, it requires users to configure their own AI service endpoints and tokens. The full-featured version is currently only available for commercial users, while the community version has limited functionality. Users need Flutter 3.7.0+ and Dart 3.7.0+ to build from source, with Android API 21+ and iOS 12.0+ as minimum requirements.

The app stands out for its mobile-optimized performance that reduces battery consumption, automatic reconnection mechanisms, mixed voice/text conversation history, and support for both cloud-based and self-hosted AI servers, making it suitable for developers and privacy-conscious users who want control over their AI infrastructure.

Xiaozhi Android Client pricing

Pricing model: Freemium

The xiaozhi-android-client itself is free and open-source under Apache-2.0 license. However, full functionality is currently only available for commercial users, while the community version has limited features. Users need to configure their own AI service endpoints which may have separate costs. For those using the official Xiaozhi server, the commercial version from Xiaoge (虾哥) now requires payment. Self-hosted xiaozhi-server can be deployed at low cost (approximately 1.5 yuan/day or 0.07 yuan/day for frontend pods). There is a separate

Xiaozhi Android Client pros

  • Cross-platform support with single Flutter codebase for iOS, Android, Web, Windows, macOS, Linux
  • Real-time WebSocket voice对话 with continuous conversation mode
  • Supports multiple AI services: Xiaozhi AI, Dify, and OpenAI with easy model switching
  • Built-in AEC+NS echo cancellation on Android for improved voice quality
  • Live2d lip-sync animation for more natural visual interaction
  • Manual call interruption and push-to-talk functionality
  • Multi-agent support allowing multiple assistants in chat list
  • Real-time speech-to-text conversion with fast response
  • Supports image messages and multimodal interaction
  • Light skeuomorphic design with smooth animations and adaptive UI
  • Automatic reconnection mechanism for stable connections
  • Voice and text conversation mixed history retention
  • Mobile-optimized performance with reduced battery consumption
  • OTA automatic device registration simplifies setup
  • Qwen3 model thinking mode toggle available
  • Open-source under Apache-2.0 license for customization
  • Supports five languages: Mandarin, Cantonese, English, Japanese, Korean
  • HTML code preview support
  • Customizable MAC address and WebSocket URL configuration
  • Dark/light theme support in development

Xiaozhi Android Client cons

  • Dark/light theme adaptation not yet implemented
  • iOS echo cancellation not fully implemented
  • Full functionality only available for commercial version
  • Community version has limited features
  • Requires own AI service endpoint configuration and token
  • Needs Flutter 3.7.0+ and Dart 3.7.0+ for building
  • iOS requires third-party sideloading without proper certificate
  • iOS users must manually enable network permissions in settings
  • No local ASR voice recognition yet
  • No local wake word detection implemented
  • No built-in TTS yet
  • No IoT mapping for phone operations
  • Limited to 502 stars indicates smaller community
  • Only 3 contributors to the project
  • 13 forks suggest limited adoption
  • Web/PC versions require self-building
  • No official pre-built binaries for all platforms
  • Commercial server from Xiaoge now charges fees
  • No official documentation in English
  • No built-in OpenAI internet search yet

Frequently asked questions about Xiaozhi Android Client

What platforms does xiaozhi-android-client support?

The app supports iOS 12.0+, Android API 21+ (Android 5.0+), Web (modern browsers), Windows, macOS, and Linux. It is built with Flutter framework allowing single codebase deployment across all these platforms.

How do I install and build the app?

Clone the repository with git clone https://github.com/TOM88812/xiaozhi-android-client.git, run flutter pub get to install dependencies, then flutter run to test. For release builds: flutter build apk --release for Android, flutter build ios --release for iOS, or flutter build web --release for Web.

What AI services can I connect to?

The client supports Xiaozhi AI service (WebSocket real-time voice conversation), Dify (with image support), and OpenAI (with text-image messages and streaming output). You can configure multiple AI service endpoints and switch between models.

Does it support echo cancellation?

Yes, it implements AEC (Acoustic Echo Cancellation) + NS (Noise Suppression) on Android devices to improve voice interaction quality. iOS echo cancellation is mentioned but not yet fully implemented, and contributors are welcome.

What languages are supported for voice recognition?

The app supports five languages for voice recognition: Mandarin (国语), Cantonese (粤语), English, Japanese, and Korean.

Is the app free to use?

The client itself is open-source and free under Apache-2.0 license. However, full functionality is currently only available for commercial users. The community version has limited features. Users also need their own AI service endpoints which may have separate costs.

How do I configure the Xiaozhi service?

You can configure multiple Xiaozhi service addresses, set WebSocket URL, provide token authentication, and customize MAC address in the service configuration settings. For official Xiaozhi service, you need a registered device MAC address.

What features are in the development plan?

Planned features include dark/light theme adaptation, support for more AI service providers, enhanced voice recognition accuracy, OTA automatic device registration, IoT mapping for phone operations, local ASR voice recognition, local wake word support, built-in TTS, MCP_Client support, and OpenAI interface with internet search.

Can I use it without a certificate on iOS?

Yes, for iOS and Mac you can use unsigned packaging methods and sideload using third-party tools. After compilation, you need to enable network permissions in Settings-APP on iOS.

What is the difference between community and commercial version?

The full-featured version is currently only available for commercial users. The community version has limited functionality. The commercial version from the developer (contact: [email protected], WeChat: Forever-Destin) provides access to all features.

Categories

Use cases

Browse all AI tools on NeedAnAI