AI-Curated Paper Notebook

A personal, agent-driven academic website for collecting, summarizing, and displaying research papers.

This repository contains both a static frontend website to beautifully showcase your reading list, and a suite of Agentic Skills that use Model Context Protocol (MCP) to autonomously discover, summarize, and link new papers.

✨ Features

• Static Frontend: Fast, lightweight HTML/CSS/JS site.
• Dynamic Filtering: Sort papers by journal, tags, or search across authors and abstracts.
• Deep Integrations:
  • 🤖 Automated Research: Uses Firecrawl to monitor specific journals (Nature, Cell, PNAS, etc.) and generate personalized summaries.
  • 📓 NotebookLM Linking: Attach your deep-dive reading notes from Google's NotebookLM directly to the paper cards.
• Agentic Engine: The repository provides predefined "Skills" (prompts + scripts) designed for an MCP-enabled AI assistant to run.

---

🚀 Getting Started

To set up your own customized paper notebook:

1. Fork this Repository

Click the Fork button at the top right of this repository to create your own copy on your GitHub account.

2. Configure Your Website

1. Open docs/index.html and update the <title> and header to reflect your name.
2. Open docs/js/app.js and edit the getBasePath() function if your repository name is different from paper-notebook.
3. Optional: Adjust the themes, colors, or CSS in docs/css/ to your liking.

3. Deploy to GitHub Pages

1. Go to your repository's Settings > Pages.
2. Under "Build and deployment", select Deploy from a branch.
3. Select the main branch and choose the /docs folder as the source.
4. Click Save. Your site will be live at https://<your-username>.github.io/<repo-name>/ in a few minutes!

---

🤖 Configuring Automated Research Skills

The real power of this notebook is the AI automation. The .agents/skills/ directory contains workflows that instruct your AI assistant on how to find and summarize papers.

Step 1: Install Required MCP Servers

To run the skills, your AI assistant must be connected to the following MCP servers:

1. Firecrawl MCP: Used for searching the web and scraping abstracts/journal tables of contents. You will need a free Firecrawl API key.
2. NotebookLM MCP: Used to fetch your personal notes and link them to papers.

Step 2: Tailor the AI's Research Interests

By default, the skills are configured to look for papers related to "plant genomics, DNA language models, and AI". You should change this to match your field!

1. Open the folders inside .agents/skills/.
2. Edit the SKILL.md file in each folder (e.g., cell_researcher/SKILL.md).
3. Find the prompt section where it describes the research interest, and replace it with your own topics (e.g., "quantum computing, superconducting qubits, and error correction").

Step 3: Run the Skills

In your MCP-enabled chat interface (like Cursor, Claude Desktop, etc.), simply tell the AI to execute a skill:

• "Run the PNAS Researcher skill to find new papers."
• "Check bioRxiv for new preprints using the bioRxiv Researcher skill."
• "Link my NotebookLM notes for the 'AlphaFold3' notebook to the website."

The AI will automatically execute the necessary tools, format the new entry, append it to docs/js/papers.json, and commit/push the changes to your repository!

---

📁 Repository Structure

├── docs/                     # Static website files served by GitHub Pages
│   ├── index.html            # Main dashboard
│   ├── paper.html            # Individual paper detail view
│   ├── css/                  # Styling
│   └── js/
│       ├── app.js            # Main rendering logic and filtering
│       ├── paper.js          # Logic for rendering the detail page
│       └── papers.json       # Database of all saved papers
└── .agents/                  # AI Automation Engine
    └── skills/               # Individual tasks the AI can perform
        ├── add_to_notebook/  # Core utility script to append to papers.json
        ├── link_notebooklm/  # Connects personal NotebookLM notes to papers
        ├── nature_journals_researcher/  # Monitors Nature family
        ├── pnas_researcher/             # Monitors PNAS
        └── ...

---

📝 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

You are free to use, modify, and distribute this codebase for your own personal knowledge management!