Documentation

The Research Memex: An Introduction

An approach to AI-powered research workflows, exploring AI as a cognitive partner.

Welcome to a New Way of Thinking About Research

Imagine having a cognitive partner that helps you navigate hundreds of papers, spot patterns across theoretical frameworks, and amplify your analytical capacity while you maintain complete intellectual ownership and develop deeper research taste.

We call this the "Research Memex." It's an approach to academic research that reimagines what becomes possible when human and machine intelligence collaborate. We're discovering how to build a new cognitive environment where AI amplifies our thinking at a scale and depth we're still exploring.

This site shares the philosophy, tools, and workflows we've developed. The core framework (interpretive orchestration, cognitive blueprints, quality control) adapts across research domains. Our implementation examples focus on systematic literature reviews, developed through our work in organizational research, though the principles transfer broadly.

Quick Start

Get up and running with the Research Memex approach

Core Philosophy

Understand the principles behind interpretive orchestration

Case Studies

See the approach applied to systematic reviews

Advanced Topics

Explore agentic workflows and MCP servers

What is a Research Memex?

In 1945, Vannevar Bush imagined the "memex," a device that would act as an intimate supplement to human memory and thought, helping us navigate and connect ideas across vast knowledge landscapes.

We're building that vision for the AI age. The Research Memex creates a space where AI becomes a genuine partner in the research process, amplifying your intellect and intuition while maintaining rigorous scholarly practice.

Core Philosophy: Interpretive Orchestration

Info

The central methodology of the Research Memex is interpretive orchestration. Rather than simply prompting AI for answers, we work as orchestrators, directing teams of specialized AI agents through complex analytical tasks.

👤 HUMAN RESEARCHER (Orchestrator)
    |
    +-- 🎯 DESIGN Workflow
    |   (Deconstruct research goals)
    |
    +-- 🎭 DIRECT Agents
    |   (Delegate specific tasks)
    |
    +-- ✓ MAINTAIN Judgment
        (Critical evaluation)
        |
        v
    🤖 AI AGENT TEAM
        - 🔍 Discovery Agent (Literature search)
        - 📊 Analysis Agent (Pattern recognition)
        - 🧠 Synthesis Agent (Theory building)
        - 🎓 Critique Agent (Quality control)
        |
        v
    🔧 RESEARCH TOOLS (Examples)
        - 📚 Reference Management (Zotero, EndNote, etc.)
        - 🕸️ Citation Discovery (Research Rabbit, Connected Papers)
        - 🤖 AI Interface (Cherry Studio, Claude Code, Gemini CLI)
        - 📝 Knowledge Base (Obsidian, Notion, Zettlr)
        |
        v
    📄 RESEARCH OUTPUT
        (Synthesis, Papers, Insights)
        |
        +-- Feedback Loop --> back to Human Researcher

Tip

Using ASCII Diagrams: You can copy this diagram directly from the code block and paste it into AI chat sessions, documentation, or text files. It's fully readable in any monospace environment and works great for explaining your research workflow to AI agents!

This approach requires deeper research thinking. As orchestrators, we:

Design the Workflow: Deconstruct complex research goals into logical sequences of analytical steps
Direct the Agents: Delegate specific cognitive tasks to appropriate AI partners
Maintain Judgment: Critically evaluate AI outputs, identify failure modes, and maintain coherent theoretical direction

The Mirror Effect: AI as a Diagnostic Partner

Tip

A key pedagogical insight of this approach is the "mirror effect." We use AI as a diagnostic partner that makes our thinking visible and, therefore, improvable.

When a vague prompt like "find gaps in the literature" yields a generic, unhelpful response, it reveals a gap in our own structured thinking. This immediate feedback loop creates deeper engagement with the material and helps us develop what we call "research taste", the expert intuition for what questions truly matter.

Case Study: Systematic Reviews

To make these concepts concrete, this site includes a detailed case study applying the Research Memex approach to conducting systematic literature reviews. This case study was originally developed for an MRes course (2025) under the guidance of Prof. Erkko Autio and Prof. Kevin Corley at Imperial Business School, Imperial College London.

Explore the Case Study →

Full acknowledgements →

Contact

Xule Lin 林徐乐
- Email: xule.lin@imperial.ac.uk
- X/Twitter: @linxule

About the Visual Identity

RM Letterforms

Seahorse Mascot

Info

Why a seahorse? The hippocampus — Greek for "horse + sea monster" — is the brain's memory center, and AIs reliably hallucinate a seahorse emoji that doesn't exist. A fitting mascot for a project about memory, AI, and the gaps between them. Read the design journey →

Built with ❤️ for researchers · Open source under MIT License