Skip to main content
Recursive Learning: The Research Memex documentation is available as an , allowing you to learn AI-powered research methodology by using AI-powered research tools. Query these docs directly from Cherry Studio, Claude Code, or any MCP-compatible AI assistant.

​
Overview

The Research Memex MCP server provides:
  • Semantic Search: Ask natural language questions about the methodology
  • Just-in-Time Help: Get setup assistance while configuring tools
  • Source Links: Every answer includes links back to relevant documentation
  • Live Updates: Always reflects the latest documentation version
MCP Server URL: https://research-memex.org/mcp

​
Quick Start

​
For Cherry Studio

  1. Open Cherry Studio → Settings → MCP Servers
  2. Click “Add MCP Server”
  3. Enter:
    • Name: Research Memex
    • Type: HTTP
    • URL: https://research-memex.org/mcp
  4. Click “Save” and “Test Connection”
  5. Try asking: “What is interpretive orchestration?”

​
For Claude Code

  1. Open settings (Cmd/Ctrl + ,)
  2. Navigate to MCP Servers
  3. Add new server configuration:
    Claude Code MCP Settings
    {
  "research-memex": {
    "type": "http",
    "url": "https://research-memex.org/mcp"
  }
}
  4. Restart Claude Code
  5. Try: “Search Research Memex for Zotero setup”

​
For Generic MCP Clients

Most MCP-compatible tools use similar configuration:
Generic MCP Client Config
{
  "mcpServers": {
    "research-memex": {
      "name": "Research Memex Documentation",
      "type": "http",
      "url": "https://research-memex.org/mcp",
      "description": "AI-powered research methodology documentation"
    }
  }
}

​
Example Queries

​
Setup & Configuration

"How do I install Zotero according to Research Memex?"
"What plugins does Research Memex recommend for Zotero?"
"Show me the Better BibTeX citation key format"

​
Conceptual Understanding

"What is interpretive orchestration?"
"Explain the mirror effect in Research Memex"
"What are cognitive blueprints?"

​
Troubleshooting

"My Zotero MCP isn't connecting"
"How do I fix API connection issues?"
"Why isn't Research Rabbit syncing with Zotero?"

​
Pedagogical Integration

​
Use Cases

First-Time MCP Users:
  • Connect to Research Memex MCP as your first server - it’s a safe, known resource
  • Test basic queries: “What is the Research Memex?” or “Show me the Quick Start Checklist”
  • Learn MCP concepts through recursive learning (using AI tools to learn about AI tools)
During Tool Setup:
  • Configuration verification: “Is my Better BibTeX setup correct?”
  • Next step guidance: “What comes after installing Zotero?”
  • Troubleshooting: “My browser connector isn’t working”
Research Workflow Support:
  • Query methodology details when needed
  • Get examples of cognitive blueprints and quality control checkpoints
  • Access failure modes and mitigation strategies
Following the Systematic Review course? This MCP is integrated throughout Sessions 2-4 as a recursive learning tool and MCP architecture example.

​
Advanced Usage

​
Multi-Query Workflows

  • Setup Verification
  • Methodology Deep Dive
  • Troubleshooting Flow
1. "Show me the Zotero setup checklist"
2. "What plugins am I missing?"
3. "How do I test my Zotero MCP server?"

​
Combining with Other MCP Servers

Use Research Memex MCP alongside:
  • Filesystem MCP: “Compare my Zotero setup to Research Memex guidelines”
  • Web Search MCP: “Search for Research Memex + academic validation”
  • Sequential Thinking MCP: “Plan my systematic review using Research Memex methodology”

​
Best Practices

Do:
  • Be specific: “How do I configure Better BibTeX citation keys?”
  • Include context: “I’m on Session 2, setting up Zotero”
  • Reference sections: “In the Zotero setup guide, what does…”
Don’t:
  • Be too vague: “Tell me about research”
  • Ask unrelated questions: “What’s the weather?”
  • Expect real-time updates: MCP reflects deployed docs, not live edits
  • Always check the source links provided
  • Cross-reference with the actual documentation
  • Understand this is search-based, not generative
  • Verify step-by-step instructions before executing
  • Use MCP for quick lookups, not deep reading
  • Still read the full documentation for comprehensive understanding
  • Ask follow-up questions to deepen comprehension
  • Document your own insights separately

​
Troubleshooting MCP Connection

“Server not responding”
  • Verify the URL is correct (check for typos)
  • Ensure you have internet connectivity
  • Try accessing https://research-memex.org in a browser first
“No results found”
  • Rephrase your query more specifically
  • Use terminology from the documentation
  • Try simpler, more direct questions
“Authentication error”
  • Research Memex MCP requires no authentication
  • Check your MCP client configuration
  • Restart your AI tool and try again

​
Why This Matters: The Meta-Learning Loop

Recursive Learning Pattern: 
THE META-LEARNING LOOP

Student needs help
    |
    v
Asks AI assistant
    |
    v
AI queries Research Memex MCP
    |
    v
Gets relevant documentation
    |
    v
Student learns method
    |
    v
Applies method using AI
    |
    v
Understands AI partnership
    |
    v
Asks BETTER questions
    |
    +─→ (loops back to AI assistant)
Students learn AI-powered research methodology by EXPERIENCING it through MCP queries - a self-reinforcing cycle where practice deepens understanding and understanding improves practice. The methodology becomes embedded through use!
The Power: You’re not just learning ABOUT AI-powered research - you’re EXPERIENCING it through the very act of accessing the documentation. The methodology becomes embedded through use.

​
MCP API Response Structure

Understanding the response format helps you work effectively with the Research Memex MCP server, especially when building custom integrations or debugging queries.

​
Search Query Response

When you query the Research Memex MCP (e.g., “What is interpretive orchestration?”), the server returns a structured response:
​
results
array
required
Array of matching documentation pages
​
metadata
object
Query metadata

​
Example Response

{
  "results": [
    {
      "title": "Core Principles",
      "url": "/introduction/core-principles",
      "excerpt": "Interpretive orchestration is the central methodology where the researcher directs specialized AI agents...",
      "score": 0.94,
      "category": "Introduction"
    },
    {
      "title": "Session 4: Advanced Agentic Workflows",
      "url": "/case-studies/systematic-reviews/session-4-agentic-workflows",
      "excerpt": "...human architect orchestrates specialized AI agents through workflow design...",
      "score": 0.87,
      "category": "Case Studies"
    }
  ],
  "metadata": {
    "total_results": 5,
    "query_time_ms": 124,
    "version": "2025-Fall"
  }
}
For Developers: Use the score field to filter low-relevance results (threshold: 0.7+). The version field ensures you’re querying the correct semester’s materials.

​
Next Steps

  1. Connect Now: Add Research Memex MCP to your AI assistant
  2. Test It: Try the example queries above
  3. Integrate: Use it during tool setup in Session 2
  4. Reflect: Notice how MCP changes your learning experience
  5. Build: Apply these lessons when creating your own MCP servers
Pro Tip: Keep a note of effective queries that worked well. These become templates for querying other MCP servers you’ll encounter.

​
Resources

⌘I