Pharos

This MCP server acts as a translator between your AI assistant and the Pharos data store. It translates your AI’s requests into Pharos GraphQL queries, fetches the results, and returns structured responses you can use in conversations about proteins, diseases, and ligands. It can be accessed as a remote HTTP endpoint or run locally via a command interface, enabling integration with chat assistants that support Model Context Protocol (MCP).

How to use

Connect your MCP-enabled AI assistant to the Pharos MCP server to enable direct questions about proteins, diseases, and chemical compounds. After connection, you can ask questions like: Pharos, tell me about the protein with UniProt ID P05067; Using Pharos, what are the known targets for Alzheimer’s Disease; or Can you use Pharos to find information on the ligand CHEMBL12. The server handles the data retrieval and returns results in a format your AI can present in chat.

Two connection methods are provided below. Choose the one that matches how your environment is set up.

How to install

Prerequisites you need before starting:

• Node.js runtime (for TypeScript-based MCP server) or the runtime your chosen MCP client supports.

Follow these steps to prepare and run the MCP server locally.

# Install dependencies if you plan to run the MCP server locally
npm install

# Build or start as defined by your local setup (example)
npm run build
npm run start

Additional content

Configuration details and usage patterns are provided below to help you integrate with your AI assistant.

Available integration methods shown here include both a remote HTTP endpoint and a local stdio (command-based) setup.

Pharos GraphQL endpoint and capabilities you can access through the MCP server include querying targets by UniProt ID, diseases by name, and ligands by identifiers like ChEMBL IDs. You can also perform GraphQL introspection to explore available fields and structures.

Available tools

pharos_graphql_query

Core tool that accepts a GraphQL query and optional variables, executes it against the Pharos GraphQL endpoint, and returns the results to your AI assistant.