claude-computer

Claude Computer is an open-source platform that demonstrates autonomous AI interaction capabilities with a CCFT token incentive system. This MCP server configuration points to a Docker-based deployment that runs the entire stack (frontend, backend, and orchestration) via Docker Compose. The frontend provides dashboards, controls, and live visualizations for observing the autonomous AI's reasoning, decisions, and remote session rendering, while the backend exposes APIs for orchestration, data handling, and real-time communication. Real-time updates are delivered over WebSocket channels, and the system supports remote session rendering through a VNC-like protocol. To get started, pull the latest release bundle, then spin up the services with Docker so you can interact with Claude’s autonomous computer interactions through the web UI at localhost:3000 (or the port defined in your config).

Once running, you can explore core capabilities such as autonomous AI reasoning and decision making, vision processing modules, prompts and tooling integrations, real-time communication, and the token incentive system. You can extend the system by adding new modules or integrations, run unit and integration tests, and contribute improvements to earn CCFT tokens. The UI exposes dashboards for monitoring AI actions, a live process viewer, and user management controls, enabling collaboration and experimentation in autonomous AI workflows.

Prerequisites:

• Docker and Docker Compose installed on your host
• Git installed on your host
• Access to the Claude Computer repository (or the release bundle zip provided by the project)

Step 1: Prepare environment

• Install Docker: https://docs.docker.com/get-docker/
• Install Docker Compose (if your Docker installation does not include it): https://docs.docker.com/compose/install/
• Ensure your user is in the docker group or run commands with sudo as needed

Step 2: Acquire the code

• Option A: Clone the repository git clone https://github.com/JeroenPieksma/claude-computer.git cd claude-computer
• Option B: Download the latest release bundle (as described in the repo notes) and extract to a working directory

Step 3: Configure environment (optional)

• If you need to customize ports or project name, edit docker-compose.yml or set environment variables as needed. For example, you can override DOCKER_HOST or COMPOSE_PROJECT_NAME in your shell or a .env file.

Step 4: Run the stack

• Start all services in development/production friendly mode using Docker Compose docker compose up --build
• Wait for the services to build and start. You should see logs indicating frontend, backend, and orchestration components coming online.

Step 5: Verify

• Open http://localhost:3000 in your browser (or the port configured in your docker-compose.yml).
• Access the UI, observe real-time updates, and test a few interactions with Claude’s autonomous platform.

Step 6: Run tests (optional)

• If the repository includes tests, run them via your preferred test runner inside the container or in a local virtual environment as described in the project documentation.

Tips and common issues:

• If you encounter port conflicts, adjust the ports in docker-compose.yml or use a .env file to override. Ensure the frontend port (default 3000) and API port do not clash with other services.
• For performance with large models or vision assets, consider increasing memory limits for Docker and CPU quotas in your docker-compose.yml.
• The token incentive system (CCFT) is a governance-enabled component; review the token economics and distribution model in the project docs to understand contribution rewards.
• If you update the frontend assets manually, ensure the build artifacts are included in the release bundle or are correctly mounted into the Docker containers.
• Check logs for WebSocket connections and remote rendering sessions if you encounter real-time or UI lags.
• Security note: expose only necessary ports and consider enabling authentication and access controls in production deployments.