CoPaw

What is CoPaw

CoPaw, also known as Co Personal Agent Workstation, is an open-source personal AI assistant workstation built on the AgentScope framework. It is designed to be deployed locally or in the cloud, offering integration with multiple messaging channels and the ability to run local Large Language Models (LLMs) with full privacy control.

How to use CoPaw

CoPaw can be installed using several methods:

• pip: For users with Python 3.10+, run pip install copaw. After installation, initialize with copaw init --defaults and launch with copaw app. The console will be accessible at http://127.0.0.1:8088/.
• curl: For macOS/Linux, use the one-line installer: curl -fsSL https://copaw.agentscope.io/install.sh | bash.
• Docker: Pull the latest image with docker pull agentscope/copaw:latest and run it using docker run -p 8088:8088 -v copaw-data:/app/working agentscope/copaw:latest.
• Cloud: One-click deployment is available on ModelScope Studio or Alibaba Cloud Computing Nest.

Features of CoPaw

• Multi-Channel Chat Integration: Natively connects with platforms like DingTalk, Feishu, QQ, Discord, and iMessage.
• Local LLM Execution: Supports running LLMs locally using llama.cpp (cross-platform) and MLX (Apple Silicon).
• Modular Agent Architecture: Features decoupled components (Prompt, Hooks, Tools, Memory) for flexibility and extensibility.
• Extensible Skill System: Includes built-in cron scheduling and auto-loading skills from the workspace.
• Long-Term Memory & Heartbeat: Proactively remembers user preferences and tasks, and can autonomously perform scheduled tasks.
• Full User Control & Privacy: Allows local deployment and choice of LLM, ensuring data privacy.
• AgentScope Framework: Built upon a developer-centric framework for building and running intelligent agents.

Use Cases of CoPaw

• Social & News Aggregation: Compiling daily digests from platforms like Xiaohongshu, Zhihu, and Reddit; summarizing videos.
• Work Productivity: Automating email digests, generating weekly reports, and extracting contacts.
• Creative Workflows: Assisting in drafting content like video scripts and social media posts.
• Research & Knowledge Management: Tracking news, building personal knowledge bases, and summarizing web information.
• Desktop Assistant: Organizing files, summarizing documents, and managing tasks via chat.
• Health & Lifestyle: Tracking diet and fitness data, setting reminders for routines.

FAQ

• What is CoPaw? CoPaw (Co Personal Agent Workstation) is an open-source personal AI assistant built on the AgentScope framework. It supports multi-channel chat, local LLM execution, and a modular agent architecture for user control and privacy.
• How do I install CoPaw? Install via pip install copaw (Python 3.10+), a curl script, Docker, or cloud deployment. After installation, run copaw init --defaults && copaw app.
• Does CoPaw support local models? Yes, CoPaw supports local LLMs via llama.cpp and MLX, eliminating the need for API keys or cloud services.
• Which chat platforms does CoPaw support? It natively supports DingTalk, Feishu, QQ, Discord, and iMessage, with the ability for developers to create custom channel plugins.
• Is CoPaw free and open-source? Yes, CoPaw is released under the Apache License 2.0 and is free to use, modify, and distribute.
• What is the relationship between CoPaw and AgentScope? CoPaw is built upon the AgentScope framework, serving as a reference implementation and key application within the ecosystem.