YouTube Dislike from Comments 作者: Dracon

Ever wonder if a YouTube video is worth your time before clicking?
YouTube Dislike from Comments analyzes the top comments using a local AI model (Twitter RoBERTa) to reveal the real audience sentiment — before you even hit play. A sleek sentiment bar appears at the bottom of every video, showing you at a glance whether viewers loved it, hated it, or felt mixed.
How it works:
1. Open any YouTube video or Short
2. Extension reads the top comments (sorted by engagement)
3. AI analyzes sentiment locally in your browser (positive, neutral, negative)
4. See the mood bar at the bottom of the page — green = loved, red = hated, gray = mixed
Key Features:
- ✅ 100% local processing — no data sent to any server
- ✅ Works offline after first model download (~50MB from HuggingFace)
- ✅ No account, no API key, no tracking
- ✅ Configurable: choose comment count (20-100), Shorts support, tooltip toggle
- ✅ Smart caching: results stored for 24 hours
- ✅ Knight Rider loading animation during analysis
Privacy First:
All comment analysis happens entirely in your browser using Transformers.js. Comment text is never transmitted. The only network request is downloading the quantized AI model on first use, which is then cached locally.
Perfect for avoiding clickbait, checking tutorial quality, or verifying product review authenticity.
YouTube removed public dislikes. This brings them back — from the voice of the community.