QwenPaw

Project Overview

QwenPaw is your personal AI assistant — deploy locally or in the cloud, extend with Skills & Plugins, connect across every channel. It never forgets: a three-layer memory system includes live working context, full verbatim history, and a self-evolving personal knowledge base powered by ReMe. Conversations and resources continuously become readable, editable, searchable, and linked Markdown memory.

Key Features

• Never forgets: Three-layer memory (live working context, full verbatim history, self-evolving personal knowledge base)
• Local or cloud, runs free: Built-in QwenPaw-Flash models (2B / 4B / 9B) trained for agent tasks, no API key needed; also works with Ollama, LM Studio, or 14+ cloud providers
• Security built in: Kernel-level Sandbox, Tool Guard, File Guard, Skill Scanner, and Access Policy. Dangerous commands are blocked before they run
• Multi-agent & parallel: Spawn independent agents with their own memory and skills. Sub-agents at runtime. Agent Communication Protocol (ACP) for cross-system orchestration
• File workspace: Unified file navigation, preview, editing, diffs, upload, and download across project and Agent files
• Extensible: Skills for scheduling, documents, browser, news, and more. Plugin architecture with a marketplace. MCP integration for external tools
• Reachable anywhere: DingTalk, Lark, WeChat, Discord, Telegram, iMessage, QQ — one instance, all channels
• Yours, not ours: Deploy locally — data stays on your machine. No third-party hosting, no data upload

Target Audience

• Users who want a personal AI assistant with full data control
• Users who need unified AI capabilities across multiple chat channels
• Users who want to automate daily tasks (news digests, report generation, multi-channel broadcasting)
• Developers and users who need code review and document processing

Quick Start

1. Visit the official documentation at qwenpaw.agentscope.io for installation options
2. Choose local or cloud deployment
3. Configure a model Provider (use built-in QwenPaw-Flash or connect Ollama / LM Studio / cloud API)
4. Connect chat channels (DingTalk, Lark, WeChat, Discord, Telegram, etc.)
5. Install Skills and plugins as needed to build custom workflows

Core Advantages

• Data sovereignty: With local deployment, data stays entirely on your machine with no third-party hosting
• Three-layer memory: Conversations and resources continuously become a searchable, linked Markdown knowledge base
• Unified multi-channel: One instance connects to all mainstream chat platforms
• Multi-agent architecture: Supports spawning independent agents, runtime sub-agents, and cross-system orchestration
• Security built in: Dangerous commands are intercepted by kernel-level Sandbox and multiple guards before execution

System Requirements

• OS: Windows, macOS, Linux
• Additional dependencies: Ollama / LM Studio required for local running, or use the built-in QwenPaw Local runtime; cloud deployment requires the corresponding platform runtime

Notes

• QwenPaw-Flash models (2B / 4B / 9B) are trained for agent tasks and suit daily assistant scenarios; for complex reasoning tasks, connecting a larger model is recommended
• For multi-user scenarios, refer to the self-hosted QwenPaw Hub solution