pytaiga

The pyTaiga MCP Bridge exposes a set of tools that map Taiga project resources into the MCP protocol. It enables AI agents to create, read, update, and delete Taiga projects, epics, user stories, tasks, issues, and sprints, while also supporting authentication flows and transport modes. By leveraging the tools and a standardized verbosity option, agents can query structured information, perform bulk operations, and manage project artifacts in a contextual, reproducible manner. The bridge supports auto-authentication via environment variables, as well as manual session control for advanced workflows. The included transport modes (stdio and SSE) let you choose between terminal-based integration and web-based, server-push interactions.

Prerequisites:

• Python 3.10 or higher
• uv package manager
• Git

Basic Installation

1. Clone the repository:

# Clone the repository
git clone https://github.com/talhaorak/pytaiga-mcp.git
cd pytaiga-mcp

2. Install dependencies via the uv workflow (as described by the project):

./install.sh

Development Installation (optional)

./install.sh --dev

Manual Installation (production dependencies only):

# Production dependencies only
uv pip install -e .

# With development dependencies
uv pip install -e ".[dev]"

Run the server locally (example)

# Default stdio transport
uv run python src/server.py

# For SSE transport
uv run python src/server.py --sse

Note: The exact paths may vary depending on your setup. Ensure that the pyTaiga MCP directory structure matches the commands above.

Environment variables are preferred for sensitive credentials to avoid exposure in logs. Common issues include:

• Wrong Taiga API URL or credentials: verify TAIGA_API_URL, TAIGA_USERNAME, and TAIGA_PASSWORD.
• Transport mode mismatches: ensure TAIGA_TRANSPORT is set to stdio or sse, or use the corresponding run.sh/flags.
• Permission errors when installing with uv: run as a user with proper permissions, or use a virtual environment.
• If using auto-auth, ensure TAIGA_USERNAME and TAIGA_PASSWORD are set before startup to enable auto-login.
• Verbosity control helps limit response size in MCP calls; use verbosity fields such as minimal, standard, or full when listing resources.