English | 中文
A comprehensive platform for managing Agent applications, helping small teams and individual users to quickly build shared Agent applications and provide cloud services for Agent applications, enabling unified management and remote asynchronous execution (e.g., Agent completes code project iterations in the cloud and automatically submits PRs).
- Asynchronous Execution of Agent Applications: Non-blocking operations, decoupling humans from Agents
- Multi-scenario Support: Adaptable to various use cases and requirements, easily extend application scenarios through custom plugins
- Private Deployment: Full control over your data flow with local deployment
The only requirement: Docker. For users with Windows OS, try to install WSL before Docker
git clone https://github.com/yjqqqaq/workmate.git
cd workmate
# Clone necessary images (due to unknown docker socket proxy bugs, will be fixed in next version)
docker pull busybox && docker pull fffiaze/backend-gitlab-runner:latest
# Start the application using docker-compose
docker-compose up --build -d- Access the web interface at
http://<host_ip>:8008. - After simple initialization setup, enter the scenario selection page.
- After selecting a scenario, fill in the necessary configuration and input for the scenario to run in the backend.
- View execution status in Docker history.
- Adding more diverse input, output, and logs support
- Adding MCP Server support (as an Agent Sidecar)
- Adding Agent Pipeline support
- Adding Agent Cluster support
Based on the current architecture design, all these features can be easily implemented within the current framework.
For detailed information on configuring scenarios, please refer to the Scenario Configuration Guide or Chinese Scenario Configuration Guide.
Special thanks to Yuan Yang for his expert technical guidance and support.
Our thanks also go to Claude-code and Docker for their outstanding work that made this project possible.