https://github.com/evansims/openfga-mcp
An experimental Model Context Protocol (MCP) server providing tools to read, search, and manipulate OpenFGA stores programmatically via Large Language Models (LLMs.)
https://github.com/evansims/openfga-mcp
Last synced: 3 months ago
JSON representation
An experimental Model Context Protocol (MCP) server providing tools to read, search, and manipulate OpenFGA stores programmatically via Large Language Models (LLMs.)
- Host: GitHub
- URL: https://github.com/evansims/openfga-mcp
- Owner: evansims
- License: apache-2.0
- Created: 2025-03-06T18:55:02.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-03-06T19:33:57.000Z (4 months ago)
- Last Synced: 2025-03-06T19:36:28.617Z (4 months ago)
- Language: Dockerfile
- Homepage:
- Size: 0 Bytes
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- mcp-index - OpenFGA MCP Server - Enables Large Language Models to interact with OpenFGA stores, facilitating reading, searching, and manipulation tasks. Supports fine-grained authorization through agentic AI and vibe coding. (Task and Project Management)
README
# OpenFGA MCP
An experimental [Model Context Protocol](https://modelcontextprotocol.io/) (MCP) server that enables Large Language Models (LLMs) to read, search, and manipulate [OpenFGA](https://openfga.dev) stores. Unlocks authorization for agentic AI, and fine-grained [vibe coding](https://en.wikipedia.org/wiki/Vibe_coding)✨ for humans.
Built using the [OpenFGA Python SDK](https://github.com/openfga/python-sdk) and [MCP Python SDK](https://github.com/modelcontextprotocol/python-sdk).
## Quick Start
### Requirements
- Python 3.10+
- OpenFGA
### Installation
```bash
# Using pip/uv
uv pip install openfga-mcp
# From source
git clone https://github.com/evansims/openfga-mcp.git
cd openfga-mcp
make setup
source activate_venv.sh
```
### Running
```bash
# Direct CLI usage
openfga-mcp-server --url "https://localhost:8000" --store "your-store-id"
# Using Make
make run
# With Docker
make docker-build
OPENFGA_API_URL="https://localhost:8000" OPENFGA_STORE_ID="your-store-id" make docker-run
```
### Connecting
Connect your LLM application to the MCP server endpoint (default: http://localhost:8090).
Compatible with [MCP clients](https://modelcontextprotocol.io/clients) including Cursor, Windsurf, Cline, Claude Desktop, and Zed.
## Development
```bash
# Setup
make setup
source activate_venv.sh
# Common tasks (all run in virtual environment automatically)
make test # Run tests
make lint # Run linting
make type-check # Run type checking
make check # Run all checks
# Interactive development
make shell # Start shell in virtual environment
make repl # Start Python REPL
make ipython # Start IPython REPL
# Run a custom command
make in-venv CMD="python -m openfga_mcp version"
```
## Use Cases
1. **Dynamic Access Control**: LLMs interpret natural language to determine permissions based on context
2. **Policy Management**: Create or adjust authorization policies through conversational interfaces
3. **Explainable Authorization**: Provide clear justifications for access decisions
4. **Policy Debugging**: Diagnose permissions issues conversationally
5. **Secure Collaboration**: Grant temporary access with precise scope
## Documentation
For detailed documentation, run:
```bash
make docs-serve
```
## Contributing
See [Contributing Guidelines](CONTRIBUTING.md) for more information.
## License
Apache License 2.0 - see the [LICENSE](LICENSE) file for details.