Your AI's research library card, in detail

Connection details

Every AI client asks for the same two values. Copy them into the connector setup form.

Per-client setup

Pick your AI client. Each takes about 2 minutes once you have your key.

The five tools

Once connected, your AI gains these capabilities. Just ask in natural language and the AI picks the right tool.

• search: 5,750+ peer-reviewed articles on RRM, NaProTechnology, FABMs, endometriosis, polyendocrine metabolic ovarian syndrome (PMOS / PCOS), and women's health. Filter by author, year, type, or tradition.
• get_article: full details on a specific article: abstract, citation, journal, year, RRM relevance score.
• find_related: articles related to one you've found, via citations and shared topics. Widen the lens around a single paper.
• check_guardrails: validate a draft against RRM editorial standards. Run before publishing anything written about RRM topics.
• check_facts: verify pregnancy rates, success percentages, and prevalence figures against the curated facts database.

For developers

Calling the MCP server programmatically from your own backend, a script or an agent framework. Use the official SDKs when you can; they handle the JSON-RPC envelope, transport, and reconnection for you.

Quick start (curl)

curl -X POST https://mcp.rrmacademy.org/mcp \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0", "id": 1, "method": "tools/call",
    "params": {
      "name": "search",
      "arguments": { "query": "endometriosis excision outcomes" }
    }
  }'

See the API reference for Python and Node.js samples plus the full request/response schemas.

Quickstart

Three steps from zero to your first authenticated call. Everything below is self-service and free.

1. Step 1: Create a free account. Visit /signup. Email and password, or sign in with Google. Free for clinicians, researchers, and developers.
2. Step 2: Generate an API key. Once signed in, open /account/mcp-keys and click Generate new key. Copy the token immediately. It is shown only once. Use it as an HTTP Authorization: Bearer YOUR_API_KEY header.
3. Step 3: Make your first request. No key is needed to reach the liveness probe. Confirm connectivity first, then call an authenticated endpoint with the curl example above.

The unauthenticated health check returns the literal string ok:

curl https://rrmacademy.org/health
# ok

API keys

Keys are self-service. Issue and revoke them yourself at /account/mcp-keys. There is no approval step and no support ticket.

• Auth model: Bearer token in the HTTP Authorization header. Free with any RRM Academy account.
• Issuance: Self-service at /account/mcp-keys. Label each key for where you use it.
• Rotation: Generate or revoke any time. Revocation is immediate; other keys keep working.

The full authentication flow, scopes, and code samples for curl, Python, and Node.js are in the API authentication guide.

Sandbox

The API includes a no-auth sandbox endpoint for the question-answering contract at POST /api/ask/sandbox. It returns a canned, deterministic response with no authentication, no rate limit, no LLM invocation, and no data mutation, so it is safe to call from any client any number of times. Use it to validate that your client correctly parses the answer shape before you integrate with the authenticated endpoint. Full details are on the Ask API and sandbox page.

curl -X POST https://rrmacademy.org/api/ask/sandbox \
  -H "Content-Type: application/json" \
  -d '{ "message": "test" }'

# {
#   "sandbox": true,
#   "answer": "Sandbox response. ...",
#   "citations": [ ... ]
# }

Every sandbox response carries an X-Sandbox: true header. Set Accept: text/event-stream for the Server-Sent Events variant, which emits the same payload followed by a [DONE] event.

Rate limits

Authenticated API responses carry standard rate-limit headers so your client can pace itself without guessing:

• RateLimit-Limit: requests allowed in the current window.
• RateLimit-Remaining: requests left in the current window.
• RateLimit-Reset: seconds until the window resets.

A 429 response means you have hit the limit. Back off and retry with exponential jitter. The no-auth sandbox is not rate limited.

Documentation

The developer portal has six pages under /connect/. This page is the first; the rest go deeper.

More

• Download the install guide (PDF): same content as this page, formatted for sharing.
• Server card: full machine-readable tool list.
• MCP descriptor: server metadata for agent discovery.
• llms.txt: site map for AI crawlers.
• Setup help: info@rrmacademy.org