Feynman Gate
He yunyan30 downloadsTurn long Markdown reading into output-first learning challenges with AI feedback.
- Overview
- Scorecard
- Updates3
🇨 中文文档
An Obsidian plugin that transforms long-form documents into an active, level-by-level learning experience — powered by the Feynman Technique and AI.
The Problem
AI-generated deep research reports and long markdown files are everywhere. But reading a 10,000-word document from top to bottom rarely leads to genuine understanding. You reach the end, close the tab, and retain maybe 20% of it.
The missing ingredient is active recall with feedback. Without being forced to articulate what you've learned, long-form reading stays passive.
What This Plugin Does
Feynman Gate sits alongside your markdown notes and turns reading into a structured, output-first process:
- Parses your document into logical learning units (by heading level).
- Guides you step-by-step — one section at a time, with a visible progress bar.
- Challenges you to explain each unit in your own words, as if teaching it to someone else.
- AI judges your explanation — not just for keywords, but for conceptual understanding — and gives feedback.
- Unlocks the next unit only when you've demonstrated sufficient grasp of the current one.
The result: a reading experience with checkpoints, accountability, and genuine comprehension gates — not just passive scrolling.
Core Features
| Feature | Description |
|---|---|
| 📄 Smart Parsing | Automatically segments any markdown file by heading hierarchy |
| 🎯 Level Selection | Choose your learning granularity: unit-level or summary-level |
| ✍️ Feynman Challenges | Explain each section in your own words before proceeding |
| 🤖 AI Evaluation | OpenRouter-compatible LLM assesses conceptual understanding |
| 📊 Progress Tracking | Persistent progress bar and session history per file |
| 🔁 Opt-in per File/Folder | Right-click any note or folder to enable learning mode |
| 🌐 Bilingual UI | Full English / Simplified Chinese interface support |
| 🎨 Theme-aware | Automatically follows Obsidian's light/dark theme |
Getting Started
Installation
Note: This plugin is currently in active development and not yet on the Obsidian community plugin registry.
Manual Installation:
- Download the latest release (
main.js,manifest.json,styles.css). - Copy the files into your vault at
.obsidian/plugins/feynman-gate/. - In Obsidian: Settings → Community Plugins → Enable
Feynman Gate.
Configuration
Go to Settings → Feynman Gate and set:
- Provider — OpenAI, DeepSeek, Qwen, OpenRouter, Moonshot, GLM, Ollama, or custom
- API Key — Your own provider API key; the plugin does not provide or pay for API access
- Base URL — Filled by the provider preset and editable for any OpenAI-compatible endpoint
- Model — The model name from your provider
- Language — English or 简体中文
The plugin uses BYOK (bring your own key). Requests go directly from Obsidian to your configured provider, and the key is stored locally in Obsidian plugin data.
Usage
- Open any markdown note in your vault.
- Right-click the file in the file explorer → "Start Feynman Learning".
- Click the 🧠 brain icon in the document header to open the learning panel.
- Choose your learning level and begin working through sections.
- Write your explanation → submit → receive AI feedback → proceed.
Roadmap
- Multi-agent review with specialized feedback dimensions
- Embedded quiz generation between sections
- Spaced repetition scheduling for revisiting weak sections
- Exportable learning summaries
Tech Stack
- Framework: Obsidian Plugin API + TypeScript
- UI: React 18
- LLM: OpenRouter API (model-agnostic)
- Persistence: Obsidian
data.jsonvia Plugin API
License
MIT