ClarifyTube 作者： samya loukili

🎓 ClarifyTube — Your Smart Visual Study Companion for YouTube

Tired of pausing complex YouTube tutorials, switching back and forth between tabs, and losing your train of thought?

ClarifyTube gives you an instant, interactive AI tutor right alongside your video without obstructing the player!



Unlike other extensions that blindly dump fixed, expensive text into the AI model, ClarifyTube puts YOU in 100% control of how much transcript dialogue and how many video frames are sent:

• ⏱ Adaptable Transcript Context Window (Your Choice, Your Budget):
• Economical Mode (30s before / 15s after): Feeds a short, compact dialogue window for quick, rapid-fire questions to minimize token consumption and maximize response speed.
• Standard Mode (60s before / 30s after): The perfect everyday balance between surrounding speech context and performance.
• Complete Lecture Mode (120s before / 60s after): Deeper surrounding context for dense mathematical proofs, code reviews, and technical lectures.
• Global Video Context: Feeds the entire video transcript to the AI for overarching lecture summaries, study outlines, and chapter breakdowns.
• Global + Local Hybrid: Macro-level video understanding combined with fine-grained local focus around your current timestamp.
• Custom Time Windows: Freely adjust the exact number of seconds before and after the timestamp directly in Settings!
• 📸 Granular Visual Frame Control:
• Single Frame (T0): Sends only the exact freeze-frame at the current timestamp — fast and token-efficient.
• Temporal Multi-Frame (T-X, T0, T+X): Sends a 3-frame sequence showing code changes, animations, or progression over time.
• Selective Frame Picker: Choose precisely which captured frame to forward to the model.
• Text-Only Mode (None): Completely disables image transmission to consume zero visual tokens when you only need transcript-based answers.

• ✏ Canvas Annotations on Frames:
  Draw arrows, circle math formulas, or write notes directly on the snapshot so the AI knows exactly what you are pointing to.
• 📎 Attach Photos from Your Computer:
  Have a textbook problem, an assignment sheet, or handwritten lecture notes? Click the paperclip button (📎) to attach it. ClarifyTube analyzes both the YouTube video and your uploaded document together!
• 📌 Revision Memos (0 Tokens Used):
  Pin important video screenshots or insert reference photos directly into your chat log. They cost 0 AI tokens, are stored locally, and appear in your exported study sheets!
• 💬 Resume Past Chats ("Continuer"):
  Revisit any saved study session from your notebooks and click "💬 Continuer" to instantly restore the entire chat history and ask follow-up questions right where you left off.
• 📓 Organized Study Notebooks:
  Group your discussions and notes by subject, course, or playlist with full search and local persistence (IndexedDB).
• 🖨 Beautiful PDF & Markdown Export:
  Download clean, illustrated revision documents with styled code blocks, rendered LaTeX math formulas, and all visual images intact.
• 🚀 Interactive Spotlight Onboarding:
  A built-in 7-step walkthrough with dynamic button highlights guides you through all tools with hands-on "Try it now!" actions.

• Google Gemini (Gemini 2.5 Flash & Pro — generous free tier via Google AI Studio)
• OpenAI (GPT-4o, GPT-4o-mini)
• Anthropic (Claude 3.5 Sonnet)
• Mistral AI & OpenRouter (100+ models)

• Your data stays yours: Your API keys, notes, discussions, and images are stored locally in your browser's IndexedDB.
• Direct connection: Questions go directly from your browser to your chosen AI provider without any intermediary servers reading your prompts.