Index / CoPaw (QwenPaw) · updated Oct 2, 2026

CoPaw (QwenPaw)

agentscope-ai/CoPaw · healthy · rank 5 of 43 by stars

A high-velocity personal AI assistant from the AgentScope team, positioned as an easy-to-install, self-hostable alternative in the OpenClaw ecosystem. It emphasizes multi-channel chat app support, a console UI, and an extensible plugin/skill architecture backed by a very active development team.

#personal-assistant#multi-channel#self-hosted#agent-framework#openclaw-alternative
Compare vs OpenClaw
GitHub ↗

Facts

Repository

Contributors242
Open issues1,007
Last commitSep 30, 2026
Release cadence~3 days
Latest releasev2.2.1

Runtime

LanguageTypeScript
Memory85 MB
Boot time180 ms
Deploymentself-hosted · cloud · desktop
Setup difficultyLow
Plugin ecosystemEmerging

Posture

LicenseApache-2.0
Local-firstYes
Cloud dependencyOptional
Multi-userYes
Privacy postureMixed

Community

Sentiment68% positive
Reddit mentions15
Web results10

Security breakdown

Composite 72 / 100 · how these are scored

Sandboxing6 / 10

higher is safer

API security7 / 10

higher is safer

Network isolation5 / 10

higher is safer

Telemetry safety6 / 10

higher is safer

Shell access risk6 / 10

higher is riskier

Model access

Read from the repository, not written by a model · 119 files examined

4 providers · runs locally · gateway support

Direct
Azure OpenAIGoogleOpenAIQwen
Gateway
LiteLLM
Local
llama.cppOllama
Compatible
OpenAI-compatible

Pinned models

gpt-5.6OpenAI

Pin last edited 11 days ago

Evidence

Decision

Why choose CoPaw (QwenPaw) over OpenClaw?

Why choose this

  • Backed by the AgentScope team with very high PR/issue velocity
  • Easy install via PyPI with polished console UI and docs site
  • Strong multi-channel chat app integration focus

Tradeoffs

  • Still in beta versioning (2.2.x beta tags), less battle-tested
  • Sandboxing and security hardening appear less mature than OpenClaw's
  • Smaller independent community outside the AgentScope ecosystem

Best fit

  • Users wanting a self-hosted personal AI assistant with multi-chat-app support
  • Teams already in the AgentScope/Qwen ecosystem
  • Developers who want an extensible plugin and skill architecture

Avoid if

  • You need a minimal, single-binary lightweight agent
  • You require mature, audited sandboxing for untrusted tool execution
  • You want a stable non-beta release cadence
Good Confidence62%

Evidence is moderately strong: README, release data, and ecosystem digests confirm active development and positioning, but Reddit matches are irrelevant noise and independent security assessments are lacking.

AI layer reviewed Sep 21, 2026 · how this is written

Star activity

35,410 stars today

Overview

CoPaw (also branded QwenPaw) is a personal AI assistant from the AgentScope team, designed to be easy to install and deploy either on your own machine or in the cloud. It supports multiple chat applications out of the box and exposes an extensible capability model via plugins, skills, and a hub gateway, distributed via PyPI with a documented console UI.

Architecturally it pairs a Python agent runtime (Python 3.11–3.13) with a TypeScript console frontend, and recent commits show heavy investment in frontend performance (lazy-loaded locales, split chat dependencies), memory/reranker configuration UI (ReMe), and stability fixes across CI, Windows support, and tool-call lifecycle handling. Ecosystem digests place it among the small set of high-velocity 'platform' projects alongside OpenClaw, processing dozens of PRs per day.

Compared to OpenClaw, CoPaw differentiates through its AgentScope/Qwen ecosystem integration, polished onboarding, and hub-gateway direction, but it remains in beta versioning with security and sandboxing hardening still catching up. It is a credible OpenClaw alternative for users who value ease of setup and multi-channel chat support over maximum isolation guarantees.

Nominate a clone

Add a new Claw

Paste a GitHub repository and tell us why it belongs on the tracker.

Opens a prefilled issue on GitHub — every nomination is public. Comfortable with a PR? Adding the repo to projects.json is faster.