================== /// MCP /// /// CHR /// ================== [server:online] [protocol:ready]
chroma-mcp
by chroma-core
A Model Context Protocol (MCP) server implementation that provides Chroma-powered vector & text search capabilities to LLM applications.
256
46
Open SourceInstallation
1. Prerequisites
• Python 3.9 or newer
• (Optional) Docker 24+
2. Clone the repository
git clone https://github.com/chroma-core/chroma-mcp.git
cd chroma-mcp
3. (Recommended) create a virtual-env
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
4. Install server dependencies
pip install -U pip wheel
pip install -e .[server] # extras name used in Chroma projects
# OR pull directly from PyPI if/when published
# pip install chroma-mcp
5. Configure the database backend (DuckDB by default)
# Environment variables
export CHROMA_DB_IMPL="duckdb+parquet" # or "postgres"
export CHROMA_SERVER_PORT=8000
# For Postgres
# export CHROMA_DB_HOST=localhost
# export CHROMA_DB_PORT=5432
# export CHROMA_DB_USER=postgres
# export CHROMA_DB_PASSWORD=yourpass
6. Run the server
python -m chroma_mcp.server # Starts FastAPI server on the configured port
Docker quick-start
docker build -t chroma-mcp .
docker run -e CHROMA_DB_IMPL=duckdb+parquet -p 8000:8000 chroma-mcp
The API will be available at http://localhost:8000 and Swagger docs at http://localhost:8000/docs.
Documentation
License: Apache License 2.0
Updated 7/30/2025