Rcsb Pdb

What Rcsb Pdb does

The Rcsb Pdb MCP server connects an MCP-compatible client to the RCSB Protein Data Bank, which contains experimentally determined macromolecular structures. It provides access to metadata for approximately 225,000 structures without requiring an API key. The service is suitable for agents that need to locate entries, inspect structure records, or retrieve details about chains, ligands, and biological assemblies.

The package is available in two forms. A hosted endpoint runs through the Pipeworx gateway, while the npm package can run as a local MCP process. Both expose the pack's RCSB PDB tools and use the same underlying data sources.

How it works

The Rcsb Pdb MCP server uses the RCSB search API for text queries and the RCSB core data API for entry details. Its hosted MCP endpoint is https://gateway.pipeworx.io/rcsb-pdb/mcp. A client can discover the available tools through MCP and call them with structured arguments.

The gateway endpoint also exposes shared Pipeworx meta-tools, so its tool list is larger than the six RCSB-specific tools. These include routing and discovery functions such as ask_pipeworx, which can send a question to another source in the wider catalog. The local stdio package exposes only the RCSB PDB tools and does not include those gateway meta-tools.

The same functions can also be called over HTTP. For example, the gateway provides a POST endpoint for the search operation and a GET endpoint for inspecting a tool definition.

Setup and configuration

For a remote MCP connection, add the hosted URL to an MCP client configuration:

{
  "mcpServers": {
    "rcsb-pdb": {
      "url": "https://gateway.pipeworx.io/rcsb-pdb/mcp"
    }
  }
}

For local execution, configure the client to run the package with npx:

{
  "mcpServers": {
    "rcsb-pdb": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-rcsb-pdb"]
    }
  }
}

The local process communicates over stdin and stdout.

Tools and capabilities

The Rcsb Pdb MCP server includes these operations:

• search(query, return_type?, limit?) searches the PDB by text.
• structure(pdb_id) retrieves an entry record with experiment, ligand, and polymer entity information.
• polymer_entity(pdb_id, entity_id) returns metadata for one polymer entity or chain.
• ligand(pdb_id, ligand_id) retrieves a ligand record.
• assembly(pdb_id, assembly_id?) retrieves a biological assembly.
• summary(pdb_id) returns a shorter view of an entry.

The underlying search endpoint is https://search.rcsb.org/rcsbsearch/v2/query, and detailed records come from https://data.rcsb.org/rest/v1/core/.

Limitations and notes

The hosted pack endpoint includes shared gateway tools in addition to the RCSB-specific operations; this cannot currently be disabled. Choose the local package if the client should see only the pack's tools. Choose the full Pipeworx gateway when direct access to all catalog packs is preferred.