Hey guys, we just want to share Open-CUAK, an open-source alternative to OpenAI's Operator product. It works out-of-the-box and you can host it locally. You can use any VLM as you like (Claude, Gemini, and of course, we support open-source models).
Give us a star on Github! https://github.com/aident-ai/open-cuak
Why Open-CUAK?
In the real world, businesses need automation that just works — agents that are reliable, scalable, and cost-effective. We believe automation should be open, accessible, and adaptable to your specific workflows. That’s why we built Open-CUAK (pronounced "quack" 🦆🗣️), as the Kubernetes for Computer Use Agents (CUA), enabling businesses to hire, teach, and manage automation agents with ease.
With Open-CUAK, you can:
âś… Run Operator-like automation workflows locally, ensuring full privacy
âś… Use vision-based automation with more flexibility and reliability, just like a human
âś… Turn any browser into an Operator-companion, with a browser extension
âś… Utilize a dedicated remote browser to mitigate risks associated, without sharing your own
âś… Use any vision-compatible model, whether frontier or open-source (Claude, Gemini, LLaVA, etc.)
âś… Bypass frustrating bot detection, unlocking more automation possibilities
🔜 Teach agents new workflows reliably, with SOP-based training
🔜 Centralize all account access in one place, managing everything agents have access to
⏳ Monitor and manage a large number of tasks, with built-in observability tools
⏳ Deploy and scale hundreds of agents to execute real-world tasks, in parallel
⏳ Open source a RL-trained CUA model to run automations, for free
Open-Source & Ready to Use
We are making agentic automation open and accessible for everyone. You can now:
- Run Open-CUAK locally and avoid expensive SaaS fees.
- Customize workflows for your industry-specific needs.
- Experiment with different vision models and grow your own automation crew.
“When automation becomes reliable, it becomes scalable. And when it becomes scalable, it becomes profitable.”
Cheers,
- Kimi