YTM (YouTube Transcripts Machine) is a web application designed to automatically extract and display timestamped transcripts from any YouTube video. It targets users who need to quickly access, review, or repurpose YouTube video content, offering a user-friendly interface with interactive timestamps and export options.

How It Works

The application leverages browser automation via the Stagehand SDK and BrowserBase. A browser session is automated to navigate to a provided YouTube URL, open the video's transcript panel, and extract individual transcript entries along with their precise timestamps. OpenAI's GPT models are utilized for the transcript extraction process, ensuring accurate data retrieval. The extracted data is then presented in a clean, interactive interface built with Next.js and React.

Quick Start & Requirements

• Primary install/run command: npm run dev
• Prerequisites: Node.js, npm, OpenAI API key, BrowserBase API key and project ID.
• Setup: Clone the repository, install dependencies (npm install), and configure environment variables by copying .example.env to .env.local and adding API keys.
• Deployment: Vercel is recommended.
• Documentation: Official Docs

Highlighted Details

• Utilizes Stagehand SDK for browser automation.
• Leverages BrowserBase for cloud execution.
• Supports OpenAI (default) and Anthropic Claude models for transcript processing.
• Offers interactive timestamps that link directly to video playback points.
• Provides options to copy transcripts to the clipboard or download as text files.

Maintenance & Community

The project is maintained by zaidmukaddam. Contributions are welcome via Pull Requests.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Permissive MIT license allows for commercial use and integration into closed-source projects.

Limitations & Caveats

The application requires API keys for OpenAI and BrowserBase, which may incur costs. The effectiveness of transcript extraction is dependent on the quality of YouTube's native transcript feature and the chosen AI model.