Evipedia MCP Server is a small server based on the Model Context Protocol (MCP), used to let AI agents query evipedia.ai. evipedia.ai is described as a continuously updated encyclopedia of evidence reviews on health and longevity interventions. The server also provides the ability to propose new interventions for review.
Purpose
The server is aimed at AI agents, enabling them to access evidence review content on evipedia.ai, including searching, browsing, and retrieving reviews, as well as submitting new intervention suggestions.
Main capabilities
The server provides the following tools:
- search_reviews(query): Returns matching reviews (by name/synonym/keyword/category), each including its URL and conclusion.
- list_reviews(): Returns the complete catalog as {topic, slug} pairs (canonical topic + slug for get_review / get_conclusion).
- list_updates(days?): Returns the update log, newest first, with each item providing {title, slug, status, date} for recently published (new) or revised (updated) reviews. With no arguments, returns the last 7 days, up to 100 items; days can be passed to get a wider time window (e.g. days: 30).
- get_conclusion(slug|url): Returns only the plain-text conclusion of the review.
- get_review(slug|url): Returns the full review as raw Markdown.
- get_metadata(slug|url): Returns structured medical metadata as JSON, including review dates (datePublished / dateModified / lastReviewed, a freshness signal not available in Markdown), typed about entities with alternative names, and an ordered reference list with PubMed PMIDs.
- suggest_review(intervention, goal?, references?, email?): Submits a new review of an intervention to evipedia's public suggestion form (the same form as on evipedia.ai/suggest).
- get_version(): Returns the package name and version of the running server.
Read tools accept either a bare slug (e.g. rapamycin) or a full evipedia.ai URL (e.g. https://evipedia.ai/rapamycin）——URL the last path segment is the slug, so URLs from search results can be passed directly.
Usage
The server has been published to npm under the package name evipedia-mcp, and runs via npx over stdio, with no global installation required. It is also listed in the official MCP Registry under the registered name io.github.forever-healthy/evipedia-mcp, so MCP-capable clients can automatically discover and install it.
For environments that cannot or do not want to perform a local installation, a hosted MCP server is also provided at https://mcp.evipedia.ai/，可通过 HTTP usage.
Client configuration:
- Local MCP (requires Node.js ≥ 18)
- Remote MCP (via HTTP)
- Claude Code: Add to the project's .mcp.json (or run claude mcp add).
- Claude Desktop: Add to claude_desktop_config.json.
- Cursor: Add to MCP settings.
The bundled /demo skill can be used to smoke-test the corresponding connection. It walks through the read tools (get_version, search_reviews, list_reviews, list_updates, get_review, get_conclusion, get_metadata) against live data from evipedia.ai.
Use cases
Suitable for scenarios that require AI agents to query evidence reviews on health and longevity interventions, retrieve review conclusions and structured metadata (including PubMed PMIDs), and submit suggestions for new interventions.
Limitations
The server is a thin client and only uses evipedia.ai's public endpoints. It does not depend on the evipedia content repository—the public interface is an API by design. It retrieves from https://evipedia.ai Real-time retrieval with a small in-process cache (both JSON indexes are very small). Mostly read-only, no authentication required. The only write path is suggest_review, which POSTs to evipedia's public suggestion form (Formspree).
Public API base URL:https://evipedia.ai
