#100100th of 118 in Agent frameworks and SDKs
cuawright
Coding-model agent that writes Playwright scripts and runs desktop tasks in a VM
- Stars
- 6.1k
- License
- MIT
- Last commit
- Oct 2026
- Language
- Python
0 on 21 Apr 20266,091 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
CUAWright gives a coding model a terminal to automate browsers and desktop applications. The browser runtime (formerly Webwright) writes and debugs Playwright scripts, with OpenAI, Anthropic and OpenRouter backends. The desktop runtime runs custom or official OSWorld-V2 tasks in an Ubuntu VM through a Responses-compatible API. It also ships Claude Code and Codex plugin manifests.
Who it is for: Engineers building browser or desktop agents, or benchmarking on OSWorld-V2
Strengths
- Browser runs leave reusable Playwright scripts, with trajectories and screenshots saved to disk
- Reports 88.1% on Online-Mind2Web and 67.9% partial score on OSWorld-V2
- Plugin manifests for Claude Code and Codex reuse the host agent, no extra API key
- Browser core is small: httpx, pydantic, playwright and typer, with three model backends
Weaknesses
- Desktop mode needs Python 3.12+, Docker and gated Hugging Face task and asset access
- Desktop runs do not export screenshot files or copy VM files to the host
- Persistent live-browser config depends on a Browserbase cloud account
- Desktop runtime is Apache-2.0 while the rest is MIT, so licensing is split
What it needs
- no GPU
- Needs Python 3.10+ (3.12+ for desktop), Playwright Chromium, Docker (desktop), Hugging Face gated datasets (desktop), Browserbase (live-browser config)
- Models: OpenAI, Anthropic, OpenRouter, Responses-compatible APIs
Quick answers
Is cuawright open source?
Yes. Its code is on GitHub under the MIT license.
Does cuawright need a GPU?
No, it runs without one.
Also in Agent frameworks and SDKs
See all 118| Rank | Project | Score |
|---|---|---|
| 1 | ppt-masterAgent skill that turns documents into natively editable PowerPoint decks | 95 out of 100 |
| 2 | openai-agents-pythonPython SDK for multi-agent workflows with handoffs, guardrails and tracing | 91 out of 100 |
| 3 | mastraTypeScript framework for building AI agents, workflows and MCP servers | 91 out of 100 |
| #4 | browser-usePython library that lets an LLM agent control a web browser | 90 out of 100 |
| #5 | copilotkitSDK for adding agent chat, generative UI and shared state to apps | 90 out of 100 |