🤖 - Reliable Automation Agents at Scale - 👥

Open CUA Kit (Computer Use Agent), or Open-CUAK (pronounced "quack" 🦆🗣️), is THE platform for managing automation agents at scale — starting with browsers. The Kubernetes for CUA agents.

🎯 Why Open-CUAK?

In the real world, for real businesses, working with real people, reliability is everything. When automation becomes reliable, it becomes scalable. And when it becomes scalable, it becomes profitable.

That’s why Open-CUAK is designed to run and manage thousands of automation agents, ensuring each one is reliable.

This project is still in its very early days, but our team is working very hard to make it a reality, soon. This is just the beginning of a new era in work, a new way to a world of abundant productivity.

And when productivity becomes truly abundant, we want to make sure it is equally distributed.

That's why we are making it open-sourced, today.

Read more in our launch blog.

---

❤️ Built and open-sourced by Aident AI team. ❤️

Quick Start

🛠️ Environment Setup

1. Make sure you have docker installed on your machine. You can download it from here.

2. Make sure you have docker-compose installed as well. Install from here.

3. Clone the repository and navigate to the root directory.

   git clone https://github.com/Aident-AI/open-cuak.git
   cd open-cuak
   

4. Install Supabase CLI.

   brew install supabase/tap/supabase
   
   # or (if you prefer `npm`)
   npm install -g supabase-cli
   

👉 Run Production Build

1. Set OpenAI API Key in .example.env file. (You can also set that in .env.production after Step 2)

   # [Required] Please add your OpenAI key
   OPENAI_API_KEY="your-openai-api-key-here"
   

2. Start the services (at repo root).

   bash quick-start.sh
   
   # or (if you have `npm` installed)
   npm run docker:up
   

3. Ta-da! It is now ready locally at http://localhost:3000.

 

Demos

1. Agent Demo #1: Canva Use Agent

   An automation agent uses Canva to create a poster for President's Day. When it encounters issues, such as whether to use Pro templates (require a Canva paid plan), it proactively asks the user for additional instructions.

   https://github.com/user-attachments/assets/f283189b-bc90-4875-8bd5-75b2a6a4bf9f

2. Agent Demo #2: Expedia Flight Search

   An agent uses built-in remote-browser running locally to search for flight tickets on Expedia.

   https://github.com/user-attachments/assets/e2dd7276-fff7-4ed1-a042-3c1a6bbecef8

3. Feature Demo: Account Management

   Account Management lets agents use your account and tools on your behalf.

   https://github.com/user-attachments/assets/1de1069d-053f-408f-b07b-61b399bfcc1f

 

Watch more on our Youtube channel, and subscribe to see more.

 

Core Features

✅ Run Operator-like automation workflows locally, ensuring full privacy
✅ Use vision-based automation with more flexibility and reliability, just like a human
✅ Turn any browser into an Operator-companion, with a browser extension
✅ Utilize a dedicated remote browser to mitigate risks associated, without sharing your own
✅ Use any vision-compatible model, whether frontier or open-source (Claude, Gemini, LLaVA, etc.)
✅ Bypass frustrating bot detection, unlocking more automation possibilities
✅ Cookie management for easy login, without the need for manual re-login
🔜 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 an RL-trained CUA model to run automations, for free

 

Development Setup

⚡ Start Development Servers

1. Install dependencies.

   # on mac
   brew install node
   
   # on linux
   sudo apt install nodejs npm
   

   npm install -g pnpm
   

   # at repo root
   # on mac
   npm run init:mac
   
   # on linux
   npm run init:linux
   

2. Start the dev servers.

   # at repo root
   npm run dev
   

3. Now, it is ready locally at http://localhost:3000.

▶️ Build and Run Production Build from Local

1. Run local production build (with docker).

   # at repo root
   npm run docker:start