Tiger Cowork
A self-hosted AI workspace unifying chat, code execution, parallel multi-agent orchestration, and project management. Each agent runs on a distinct provider — OpenAI-compatible APIs, Claude Code, or Codex CLI. MCP servers extend the toolbox. Built fo
What is Tiger Cowork?
Tiger Cowork is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to self-hosted ai workspace unifying chat, code execution, parallel multi-agent orchestration, and project management. each agent runs on a distinct provider — openai-compatible apis, claude code, or cod...
A self-hosted AI workspace unifying chat, code execution, parallel multi-agent orchestration, and project management. Each agent runs on a distinct provider — OpenAI-compatible APIs, Claude Code, or Codex CLI. MCP servers extend the toolbox. Built fo
This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- A self-hosted AI workspace unifying chat, code execution, pa
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx tiger-coworkConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Tiger Cowork
Tiger Cowork is a self-hosted AI workspace that unifies chat, code execution, parallel multi-agent orchestration, and project management behind a single web interface running on localhost. Each agent can be backed by a different provider — OpenAI-compatible APIs, Claude, or Codex CLI — and the workspace supports MCP servers as external tool extensions. With 16 built-in tools covering web search, Python execution, React rendering, shell commands, file operations, and seven sub-agent orchestration topologies, Tiger Cowork is designed for developers who want an open, customizable alternative to managed AI workspaces with full control over their data and models.
Prerequisites
- Docker Desktop installed (for the one-click installers) OR Node.js 18 or higher with npm
- Python 3 (optional, for Python execution features)
- API keys for at least one LLM provider (OpenAI-compatible, Anthropic, or Codex)
- A ClawHub token (prompted during first-time setup)
- An MCP-compatible client or browser access to http://localhost:3001
Install Tiger Cowork
On macOS or Linux, use the one-line installer. On Windows, use the PowerShell equivalent. Both require Docker Desktop. Alternatively, clone and build manually with Node.js.
# macOS/Linux
curl -fsSL https://raw.githubusercontent.com/Sompote/tiger_cowork/main/install.sh | bash
# Windows PowerShell
irm https://raw.githubusercontent.com/Sompote/tiger_cowork/main/install.ps1 | iex
# Manual
git clone https://github.com/Sompote/tiger_cowork.git
cd tiger_cowork && bash setup.shBuild and start the server
If installing manually, build and start the Node.js server. The workspace will be available at http://localhost:3001.
npm run build && npm startSecure your installation
Before adding API keys, change the default authentication token in the .env file. The default value 'your-secret-token-here' must be replaced before exposing the service to any network.
Add an LLM provider
Open http://localhost:3001 in your browser, go to Settings, add your API key, model endpoint URL, and select a model. Click 'Test Connection' to verify.
Optionally run as a persistent background service with PM2
For long-running sessions, use PM2 to keep Tiger Cowork alive across terminal restarts.
npm install -g pm2
npm run build
pm2 start npm --name "tiger-cowork" -- start
pm2 startup && pm2 saveTiger Cowork Examples
Client configuration
Tiger Cowork runs as a standalone web app and MCP host. To register it as an MCP server in Claude Desktop, point at the local server with your token.
{
"mcpServers": {
"tiger-cowork": {
"command": "npx",
"args": ["tiger-cowork"],
"env": {
"TIGER_COWORK_TOKEN": "your-secret-token-here",
"TIGER_COWORK_PORT": "3001"
}
}
}
}Prompts to try
Once Tiger Cowork is running, use these prompts in the web interface or via a connected MCP client to orchestrate multi-agent workflows.
- "Search the web for the latest papers on diffusion models and summarize the top 3"
- "Execute a Python script that generates a bar chart of my project's file sizes"
- "Spin up three parallel agents: one to write tests, one to write docs, one to refactor this function"
- "Create a new project task: implement OAuth login, assigned to the code agent"
- "Load the filesystem MCP server and list all files in my project directory"Troubleshooting Tiger Cowork
Server fails to start or port 3001 is already in use
Check if another process is using port 3001 with 'lsof -i :3001' on macOS/Linux. Set a different port in the .env file (e.g. PORT=3002) and restart the server.
LLM provider test connection fails
Verify the API key is correct and the model endpoint URL is reachable. For OpenAI-compatible providers, ensure the base URL ends in /v1 (e.g. https://api.openai.com/v1). Check your network for proxy or firewall restrictions.
AI agent executes unexpected shell commands or deletes files
Tiger Cowork agents can execute arbitrary code. Per the project's security warning, run only in Docker or another sandboxed environment on a machine with no sensitive data. Do not expose port 3001 to the internet without authentication.
Frequently Asked Questions about Tiger Cowork
What is Tiger Cowork?
Tiger Cowork is a Model Context Protocol (MCP) server that self-hosted ai workspace unifying chat, code execution, parallel multi-agent orchestration, and project management. each agent runs on a distinct provider — openai-compatible apis, claude code, or codex cli. mcp servers extend the toolbox. built fo It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Tiger Cowork?
Follow the installation instructions on the Tiger Cowork GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Tiger Cowork?
Tiger Cowork works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Tiger Cowork free to use?
Yes, Tiger Cowork is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Tiger Cowork Alternatives — Similar Coding Agents Servers
Looking for alternatives to Tiger Cowork? Here are other popular coding agents servers you can use with Claude, Cursor, and VS Code.
Dify
★ 142.2kProduction-ready platform for agentic workflow development.
Ruflo
★ 54.0k🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, self-learning swarm intelligence, RAG integrat
Goose
★ 45.7kan open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
Antigravity Awesome Skills
★ 38.3kInstallable GitHub library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and more. Includes installer CLI, bundles, workflows, and official/community skill collections.
AgentScope
★ 25.5kBuild and run agents you can see, understand and trust.
Serena
★ 24.5kA coding agent toolkit that provides IDE-like semantic code retrieval and editing tools, enabling LLMs to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements.
Browse More Coding Agents MCP Servers
Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Tiger Cowork in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Tiger Cowork?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.