# PubMed MCP server

> Search literature and retrieve structured citation metadata.

- Provider: PubMed
- Category: Research (https://mcpheaven.com/catalog/category/research)
- Status: Planned
- Slug: `pubmed`
- Verified: 2026-11-10
- JSON: https://mcpheaven.com/api/v1/capabilities/pubmed
- HTML: https://mcpheaven.com/catalog/pubmed

## Why it is in the catalog

Important biomedical research source whose MCP connectors are community maintained.

## Questions

### Is there an official PubMed MCP server?

Important biomedical research source whose MCP connectors are community maintained. mcpheaven lists PubMed so agents can find it by task and reach it through one MCP endpoint.

### What can an AI agent do with PubMed?

Search literature and retrieve structured citation metadata. Agents find it by describing the task, for example with the search_capabilities tool.

### How do I connect PubMed to Claude, Codex or Cursor?

Add mcpheaven once as an MCP server (https://mcpheaven.com/mcp) with an API key, then ask your agent for the task. It calls get_capability with the slug "pubmed" for details.

### Is the PubMed integration available?

The PubMed integration is planned. Agents can already discover it through mcpheaven's search, and it will become callable once the integration ships.

## Connect

MCP endpoint: `https://mcpheaven.com/mcp` (Streamable HTTP, JSON-RPC). Without a key: public catalog, 120 calls per hour per address. With a key (`Authorization: Bearer mch_…`, free at https://mcpheaven.com/signup): plan quota and private capabilities.

```bash
claude mcp add --transport http mcpheaven https://mcpheaven.com/mcp --header "Authorization: Bearer mch_…"
```

```json
{ "mcpServers": { "mcpheaven": { "url": "https://mcpheaven.com/mcp", "headers": { "Authorization": "Bearer mch_…" } } } }
```

REST: `GET https://mcpheaven.com/api/v1/capabilities?q=<task>` · OpenAPI: https://mcpheaven.com/openapi.json · Server card: https://mcpheaven.com/.well-known/mcp.json
