Let AI truly control your Android phone.
Architecture aligned with OpenClaw (280k+ Stars), bringing full AI Agent capabilities to your phone — see the screen, tap apps, run code, connect platforms.
Docs · Quick Start · Community
WeChat, Alipay, TikTok, Taobao, Maps... Anything you can do manually, AI can do too.
You: Open WeChat and send "See you tomorrow" to John
AI: → Open WeChat → Search John → Type message → Send ✅
You: Got an address in WeChat, navigate me there
AI: → Copy address from WeChat → Open Maps → Search → Start navigation
Execute commands via Termux SSH (Shell works directly, Python/Node.js need to be installed in Termux):
You: Use Python to analyze the CSV in my Downloads folder
AI: → exec("python3 analyze.py") → Return analysis results
You: Search for today's tech news
AI: → web_search("tech news") → Return titles + links + summaries
Control your phone AI remotely via Feishu, Discord, and more:
| Channel | Status |
|---|---|
| Feishu | ✅ Available (WebSocket real-time, DM/group, 39 Feishu tools) |
| Discord | ✅ Available (Gateway v10, DM/group, permission policies) |
| In-app Chat | ✅ Available |
| Telegram | 🔧 In Development |
| Slack | 🔧 In Development |
Download from the Release page:
| APK | Description | Required? |
|---|---|---|
| AndroidForClaw | Main app (Accessibility Service, Agent, Gateway) | ✅ Required |
| BrowserForClaw | AI Browser (web automation) | Optional |
- Install — Download and install AndroidForClaw
- Configure — Open the app, enter an API Key, enable Accessibility + Screen Capture permissions
- Chat — Talk directly in the app, or send messages via Feishu/Discord
846 Kotlin source files · 167,000+ lines of code · 10 modules
| Feature | Description |
|---|---|
| Playwright Mode | Screen ops aligned with Playwright |
| Context Management | 4-layer protection |
| Multi-model | Qwen 3.6 Plus · DeepSeek R1 · Claude Sonnet 4 · Gemini 2.5 |
| MCP Server | Expose accessibility/screenshot to external agents (port 8399) |
| Skill System | Built-in Skills + ClawHub online installation |
| Tool | Function |
|---|---|
device |
Screen ops: snapshot / tap / type / scroll / press |
read_file |
Read file contents |
write_file |
Create or overwrite files |
edit_file |
Precise file editing |
list_dir |
List directory contents |
exec |
Execute commands |
web_search |
Brave search engine |
web_fetch |
Fetch web page content |
javascript |
Execute JavaScript |
tts |
Text-to-speech |
| Provider | Models | Notes |
|---|---|---|
| OpenRouter | Qwen 3.6 Plus, Hunter Alpha, DeepSeek R1, Claude Sonnet 4 | Recommended |
| Gemini 2.5 Pro, Gemini 2.5 Flash | Direct | |
| Anthropic | Claude Sonnet 4, Claude Opus 4 | Direct |
| OpenAI | GPT-4.1, o3 | Direct |
| Custom | Any OpenAI-compatible API | Ollama, vLLM, etc. |
git clone https://github.com/SelectXn00b/AndroidForClaw.git
cd AndroidForClaw
./gradlew assembleRelease| Project | Description |
|---|---|
| OpenClaw | AI Agent framework (Desktop) |
| iOSForClaw | OpenClaw iOS client |
| AndroidForClaw | OpenClaw Android client (this project) |
MIT — LICENSE
⭐ If this project helps you, please give it a Star! ⭐