Video Chapters Generator

Free • Non-commercial • Open Source

Organize YouTube videos into chapters with AI-powered headlines and timecode generation.

Features

One-Click Chapter Generation: Generate chapters directly from YouTube video pages
Multiple AI Providers: Google Gemini (direct API) or OpenRouter (DeepSeek R1, Claude, GPT-4o, Llama, and more)
Free Model Options: Use free models with minimal setup (API key required for authentication but no usage costs)
Custom Instructions: Add personalized instructions to tailor chapter generation to your needs
Direct Video Navigation: Jump directly to any chapter timestamp from the results page
Cross-Browser Support: Works on both Chrome (Manifest V3) and Firefox (Manifest V2)
Multi-Language Interface: Localized in 12 languages (English, Spanish, French, German, Italian, Russian, Ukrainian, Hebrew, Chinese Simplified, Portuguese, Arabic, Korean)
Session-Based Results: Results are stored only for the current browser session
Open Source: Full transparency with code available on GitHub

Perfect For

Content creators, vloggers and streamers organizing their videos
Students creating study guides from educational content
Researchers analyzing video content structure
Podcasters converting video episodes to chapters
Anyone wanting to make videos more navigable
Educators creating lesson breakdowns

Installation

Chrome Web Store

To be published

Firefox Web Store

https://addons.mozilla.org/en-US/firefox/addon/video-chapters-generator/

Firefox Add-on

Download the latest Firefox add-on (.xpi file) from the GitHub Releases section and install it directly in Firefox.

Manual Installation

For developers or advanced users, see DEVELOPMENT.md for manual installation instructions.

Setup

Model Selection & API Configuration

The extension supports multiple AI providers with different models:

Free Options (API Key Required for Authentication)

DeepSeek R1 Free: Latest reasoning model, no usage costs but requires OpenRouter API key
Multiple Free Models via OpenRouter: Llama, GLM, and other models with $0 usage cost

Gemini Models (Direct Google API)

Gemini 2.5 Pro: Most capable model for complex analysis
Gemini 2.5 Flash: Faster model optimized for speed

OpenRouter Models (Single API Key for Multiple Providers)

DeepSeek R1: Advanced reasoning capabilities
Claude 3.5 Sonnet/Haiku: Anthropic's powerful models
GPT-4o/GPT-4o Mini: OpenAI's latest models
Llama 3.3 70B: Meta's advanced open-source model
Gemini via OpenRouter: Access Gemini through OpenRouter

API Key Setup

Option 1: Use Free Models (via OpenRouter)

Create a free account at OpenRouter
Generate an API key from your OpenRouter dashboard
Click the extension icon and select any free model (e.g., "DeepSeek R1 0528 (Free)")
Enter your OpenRouter API key
Generate chapters with no usage costs!

Option 2: Gemini Direct API

Get a free Gemini API key from Google AI Studio
Note: Gemini's free tier is sufficient for most users - no paid plan required
Availability: Gemini API is available in most countries. Users in supported regions should register for the free tier, which provides generous usage limits for chapter generation
Open extension options page and enter your Gemini API key
Select any Gemini model in the popup

Option 3: OpenRouter (Paid Models)

Create a free account at OpenRouter
Generate an API key from your OpenRouter dashboard
Open extension options page and enter your OpenRouter API key
Select any OpenRouter model in the popup

Note: API keys are stored securely in your browser's local storage and never leave your device except when sent to the respective AI service.

Cost Information: The extension itself is completely free. AI model usage costs (if any) are paid directly to AI providers - no money goes to the extension developer. Most users can operate entirely within free tiers.

Disclaimer: No warranty is provided for external APIs and services. Free AI models may have usage limitations, rate limits, or availability restrictions that are beyond our control.

Usage

Generating Chapters

Open any YouTube video with captions
Click the extension icon
Select your preferred AI model
Enter your API key (required for all models - free models have no usage costs)
Add custom instructions to guide chapter generation (optional)
Click "Generate Chapters"
Copy formatted chapters from the results page

Example Custom Instructions

You can also try running without any custom instructions to see the default AI behavior. Use the extension button on the results page to easily modify instructions or change models for the same video.

"Headlines should be brief yet meaningful, reflect the content, and match the individuality of the video participants. Additionally write a brief summary"
"Headlines should be detailed, reflect the content, and match the individuality of the video participants"
"Write both the headline and one-sentence thesis for each without adding prefixes like 'Thesis:'"
"For each headline, provide both thesis and counter-argument"

Instruction History

Save frequently used instructions for easy reuse
Click on saved instructions to automatically fill the input field
Delete instructions you no longer need

Results Management

View Results: Opens results in a new tab (or focuses existing tab for the same generation)
Back to Video: Smart navigation that returns to the original video tab
Copy Chapters: Copy generated chapters with video URL included

Privacy & Security

Data Transmission: The extension sends YouTube video transcripts to your selected AI provider for chapter generation:
- Google Gemini models: Data sent directly to Google's Gemini API (ai.google.dev)
- OpenRouter models: Data sent to OpenRouter API (openrouter.ai), which forwards to the selected model provider (Anthropic, OpenAI, Meta, DeepSeek, etc.)
- User Control: You explicitly choose which AI service processes your data by selecting the model
Local Storage Only: All your settings and results are stored solely in your browser's local storage
No Extension Server: The extension does not operate its own servers or collect any user data
Transcript Retrieval: Subtitles are retrieved from YouTube using the same method the browser uses when users open the transcript panel
Minimal Permissions: The extension requests only the permissions necessary for YouTube integration and session storage
Session-Only Results: Generated chapters are stored only for the current browser session and are not persisted
API Key Security: API keys are stored locally in your browser and only transmitted to their respective AI services
Localized Interface: Extension interface automatically adapts to your browser language (supports 12 languages)

License

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

It is licensed under the GNU General Public License v3.0 or later.

Recent Updates

v2.0.0 - Multi-Provider AI Support

🚀 New Providers: Added OpenRouter integration with 8+ AI models
🆓 Free Options: DeepSeek R1 and other models with no usage costs (API key required for authentication)
🎯 Model Selection: Choose from Gemini, Claude, GPT-4o, Llama, and more
🏗️ Modular Architecture: Clean separation of AI providers with shared functionality
🔧 Smart UI: Dynamic API key field based on selected model
📊 Provider Info: Results page shows which model and custom instructions were used

Acknowledgments

Subtitle retrieval logic is based on code by Hamza Wasim, used with permission.

Chrome Web Store Distribution

This extension is distributed under GPL v3.0 license, which is compatible with Chrome Web Store distribution. The extension:

Contains proper copyright notices and attribution
Includes complete source code access via GitHub repository
Data Transmission Disclosure: Sends YouTube video transcripts to user-selected AI providers (Google Gemini API or OpenRouter API) for chapter generation only
Uses established YouTube transcript extraction methods with author permission
Maintains user privacy with local-only settings and session storage