copaw was rebranded to qwenpaw (new pip package name is qwenpaw).

I wanted to keep everything working the same inside my venv, so I did:

1. Activate venv

source venv/bin/activate

2. Install new package

pip install -U qwenpaw

3. Remove old one

pip uninstall copaw

4. Symlink inside the venv bin

ln -s qwenpaw venv/bin/copaw

5. Quick check

copaw --help

Also, qwenpaw still picks up existing agent configs from ~/.copaw, so no changes needed there for now.

Didn’t have to touch any scripts  copaw still works 👍

CoPaw v1.0.0 is here with full Multi-Agent support

Now you can run multiple agents, each with its own brain, memory, skills, and tools, and even let them collaborate with each other

What’s supported?

Multi-Agent Workspace create and manage multiple agents, each with its own configuration, memory, skills, and conversation history

Inter-Agent Collaboration Have agents communicate with each other to accomplish complex tasks together

Perfect for separating roles like coding, research, writing, etc.

⚙️ How to Set It Up

1. Update CoPaw

If you're on an older version, run:

copaw update

This will automatically migrate your existing workspace and configs to the new multi-agent setup.

Manage Agents via UI Console

https://preview.redd.it/bfqgui8y0csg1.png?width=3524&format=png&auto=webp&s=9ea18c52835e8db0a74de711798a0b622cf7ca12

https://preview.redd.it/4izvmjxw0csg1.png?width=1288&format=png&auto=webp&s=c8b05a3a2d668835e1b1d673fecc71e7bfbbc7ea

https://preview.redd.it/ag9imyvv0csg1.png?width=1498&format=png&auto=webp&s=834db1077712f21ee34ef7b7f5b08c11a517ea8a

Enable Inter-Agent Communication

• Turn on the Multi-Agent Collaboration skill
• prompt agents to collaborate eg "Ask the writing agent to write a blog post on ABC".

CoPaw v0.2.0 just dropped - agents talking to each other, a built-in QA bot, and stability upgrades

This is a pretty substantial update here's what's new:

New features

• Inter-Agent Communication - new copaw agents and copaw message CLI commands let agents list each other, push messages to channels, and send requests between agents
• Built-in QA Agent - ships pre-configured to answer CoPaw installation and usage questions out of the box
• Configurable LLM Auto-Retry - retry behavior is now per-agent and configurable from the console Settings page
• File Access Guard - configurable deny list for sensitive files and directories
• Audio/video & voice input in the console - media support has landed in the main UI
• Stream reconnect after page refresh - streaming conversations now resume properly if you refresh
• Model provider search - you can now search across providers in the console
• Multimodal capability detection - the UI now detects and labels which models support multimodal input

Improvements

• Custom working directory - COPAW_WORKING_DIR now works correctly throughout the codebase (hardcoded ~/.copaw paths are gone)
• Dynamic token counting - now uses per-agent config instead of a single global tokenizer
• Config auto-repair - if config.json gets corrupted, CoPaw automatically repairs it or backs it up and starts fresh instead of crashing
• Enhanced grep and glob search tools
• Summarization improvements - when an agent hits max iterations, a "round ended" notice is shown to guide the user
• Provider/model sync - the console now re-fetches active models on navigation so the UI stays accurate