ChEMBL

The Unofficial ChEMBL MCP Server exposes a comprehensive set of tools that give AI assistants and MCP clients direct programmatic access to ChEMBL data through a stable MCP interface. The server runs as an MCP server that communicates via stdio and exposes 22 specialized tools spanning core chemical search, target analysis, bioactivity data, drug development, chemical property analysis, and advanced search capabilities. Typical usage involves starting the server, then configuring your MCP client (or CLI) to connect to the server and invoking any of the available tools with the appropriate arguments. Each tool accepts a defined set of parameters (for example, query strings for searches, identifiers like CHEMBL IDs, and optional filters) and returns structured results that can be consumed by your AI assistant or downstream workflows.

To use it in your MCP client, add the server to your MCP configuration with its command and arguments, then call tools by name (e.g., search_compounds, get_compound_info, search_targets, etc.). The server also provides resource URI templates to facilitate direct integration with ChEMBL data, enabling seamless linking from your application to compound, target, assay, and activity information. Remember to supply environment variables if your deployment requires API keys or other credentials, though the default setup in the repository typically uses public ChEMBL REST endpoints.

Prerequisites:

• Node.js v16 or higher
• npm (or yarn)
• Git

Setup steps:

1. Clone the repository:

git clone <repository-url>
dcd chembl-server

2. Install dependencies:

npm install

3. Build the project (if required by the project structure):

npm run build

4. Start the MCP server:

npm start

5. (Optional) Run via Docker:

# Build the image
docker build -t chembl-mcp-server .

# Run the container
docker run -i chembl-mcp-server

Integrate with an MCP client by adding the server to your client configuration, pointing to the appropriate command and arguments as shown in the examples below.

Notes and tips:

• The server exposes 22 tools; refer to the Tools section in the README for exact names and parameter schemas.
• If running behind a private network or requiring API authentication, configure environment variables as needed (e.g., API keys, base URLs).
• When using Docker, you can bind-mount volumes for persistent configuration or data if needed by your deployment workflow.
• If you encounter timeouts or API rate limits from ChEMBL, consider adding reasonable limits and retry logic on the client-side, or caching frequent queries.
• Ensure you are using Node.js v16+ for compatibility with modern dependencies.