Video to Text Assistant | YouTube/Bilibili Text Extractor

🚀 Core Functions 1. Video to Text Utilizing advanced Whisper AI models for precise speech recognition in videos. Supports multilingual recognition (90+ languages including Chinese, English, Japanese, etc.). Automatically generates complete transcripts, downloadable as TXT files. 2. Supports Mainstream Video Platforms ✅ YouTube (including short links youtu.be) ✅ Bilibili 3. Local Processing 100% Local Execution: All audio and video processing is done on your device. Zero Data Upload: No video, audio, or text content is sent to external servers. Offline Availability: Once initialized, it works even without an internet connection. 🔒 Privacy & Security We understand the importance of privacy, hence the following design: Cookies for Local Download Only: Collected cookies are used solely by the local yt-dlp tool to access videos you are logged into (e.g., age-restricted content), and are never sent to any server. No Tracking, No Ads: We do not collect any user behavior data nor insert any ads. Open Source & Transparent: Source code is fully open on GitHub, welcoming review and contribution. 💡 Key Features ✨ Free & Unlimited - Completely free with no limits on transcription count or duration. ⚡ Efficient & Fast - Utilizes your device's CPU/GPU resources with intelligent threading for ultra-fast transcription. 🎯 Accurate Recognition - Based on OpenAI Whisper model for high accuracy. 🔧 Easy to Use - Open the sidebar, click "Create Task" to start, no complex configuration needed. 📦 Lightweight Installation - Compact browser extension with an optimized Native Host installer. 📋 How to Use Step 1: Install Extension Install this extension from the Chrome Web Store. Step 2: Install Native Host (Required) Click the extension icon to verify the connection. If not installed, a welcome page will guide you. Download the Native Host installer for your OS (Windows / macOS). Run the install script and wait for completion. Refresh the extension to start using. Step 3: Start Transcribing Open any YouTube or Bilibili video page. Click the extension icon in the browser toolbar to open the sidebar. Click the "Create Task" button. The extension will automatically download audio and start transcribing. Once done, click "Download Result" to get the TXT transcript. ⚙️ System Requirements Browser: Google Chrome or Chromium-based browsers (Edge, Arc, etc.) Version requirement: 88 or above Operating System: Windows: Windows 10/11 (64-bit) macOS: macOS 11 Big Sur or above (ARM64 recommended for best performance) Hardware Recommendations: Processor: At least 2-core CPU (4-core+ recommended) RAM: At least 4GB RAM (8GB+ recommended) Disk: At least 2GB free space (for model files and temporary audio) ❓ FAQ Q: Why is the Native Host required? A: Due to browser security restrictions, large AI models cannot run directly in the extension. The Native Host is a local service program that handles audio downloading and AI transcription, ensuring everything runs safely on your device. Q: How long does transcription take? A: It depends on video length and your device performance. Typically, a 10-minute video takes about 2-5 minutes to transcribe (faster on better hardware). Q: Which languages are supported? A: The Whisper model supports 90+ languages, including but not limited to Chinese (Mandarin, Cantonese), English, Japanese, Korean, French, German, Spanish, etc. The model automatically detects the language. Q: How to uninstall? A: After removing the extension from Chrome, you also need to manually uninstall the Native Host. Please download the uninstall script from our GitHub Release page and run it. =========================================== 📝 Version History =========================================== 【v1.0.5】(2026-01-31) • 🚀 Background Tasks - Tasks continue running after sidebar closes; notifications sent upon completion. • 🧪 Stability Improved - Auto-reconnect after MV3 Worker restarts. • 🔧 Windows Optimization - Improved install/uninstall scripts; fixed path issues. • 📚 Documentation - Added detailed troubleshooting guide. 【v1.0.4】(2026-01-30) • 🚀 Bundled Node.js - Solves YouTube signature decryption; no manual Node install needed. • ⚡️ Smart Fallback - Auto-switches to mobile interface if Node environment is unavailable. • 🐛 Bug Fixes - Fixed Windows Native Host startup issues. 【v1.0.3】(2026-01-24) • 📦 Automated Build - Introduced GitHub Actions for auto-packaging releases. • 🔄 Auto-Update yt-dlp - Automatically integrates the latest downloader with every release. • 📝 Smart Changelog - Auto-extracts release notes from README. 【v1.0.1】(2026-01-15) ✨ New Features • Added manual native service re-check function. • Automatically disable "Add Task" button when service is not installed. • Added "Recheck" button in the installation guide panel. 🚀 Performance • Removed 10s startup delay; closes overlay immediately when ready. • Optimized detection flow with early checks. • Fixed overlay control logic. 💡 UX Improvements • Smarter onboarding timing. • Simplified status indicator. • Clearer installation prompts and error feedback. 🐛 Bug Fixes • Fixed first-launch connection issues. • Fixed overlay closing logic. • Cleaned up console logs. 【v1.0.0】(2026-01-14) • Initial Release • Basic video-to-text functionality • Support for YouTube and Bilibili • Local AI transcription (Faster-Whisper) =========================================== 🔗 Get Help GitHub Repo: https://github.com/kangchainx/video-text-chrome-extension Issues: https://github.com/kangchainx/video-text-chrome-extension/issues Releases: https://github.com/kangchainx/video-text-chrome-extension/releases 📜 License & Credits Open sourced under MIT License. Thanks to: OpenAI Whisper (faster-whisper implementation) yt-dlp (Video downloading) FFmpeg (Audio processing) Note: On first use, the AI model file (~140MB) will be downloaded automatically. Please ensure internet access and wait patiently. Downloaded once, used forever.