A Model Context Protocol (MCP) server that provides AI assistants with programmatic access to the Avni platform. This server enables LLMs like OpenAI GPT-4 to interact with Avni's system through standardized tool calling.
- Built-in monitoring support at
/health - Deployed with HTTP transport and proper error handling
- Python 3.13+
- uv package manager
# macOS
brew install uv
# Other platforms
# See: https://docs.astral.sh/uv/getting-started/installation/git clone https://github.com/avniproject/avni-mcp-server
cd avni-mcp-server
uv syncCreate a .env file in the project root:
# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key_here
# Avni Platform Configuration
AVNI_BASE_URL=https://staging.avniproject.org
AVNI_AI_SERVER_URL=https://staging-mcp.avniproject.org/
# Server Configuration
PORT=8023Staging:
AVNI_BASE_URL=https://staging.avniproject.org
AVNI_AI_SERVER_URL=https://staging-mcp.avniproject.org/# Run the server
uv run avni-mcp-server
# Server will start on http://localhost:8023
# MCP endpoint: http://localhost:8023/mcp (no longer used, as we call tools in code now)
# Health check: http://localhost:8023/health# Run linting
uv run ruff format .
# Run tests
uv run pytest
# Run with coverage
uv run pytest --cov=.
# Run specific test
uv run pytest tests/test_tools.pyEnsure your deployment exports these variables:
export OPENAI_API_KEY={{ openai_api_key }}
export AVNI_BASE_URL={{ avni_base_url }}
export AVNI_MCP_SERVER_URL={{ avni_mcp_server_url }}
export PORT=8023# Check server logs
sudo journalctl -u avni-ai -f
# Development debugging
uv run python main.py --log-level DEBUG