StudyPilot ( Learning Mode for the Web )

StudyPilot is a Chrome extension that turns any article into tidy study notes with on-device AI. It summarizes sections, creates flashcards, answers questions from the page, explains selected text, and can analyze diagrams all without sending your data to a server.

Built for the Google Chrome Built-in AI Challenge 2025.

---

Features

• Section aware summaries
  Detects page sections (H1–H3) and produces concise, markdown bullet points per section.

• Multilingual output
  Choose your language (English, Hindi, Spanish, French). Translation uses the on device Translator API when available, with a Prompt API fallback.

• Flashcards in one click
  Generates up to 8 Q↔A cards as JSON using the Prompt API, with CSV export.

• Ask about this page
  Short, sourced answers using only the extracted page text (no hallucinated facts).

• Diagram Q&A
  Upload an image/diagram and get a step by step explanation (multimodal Prompt API).

• Explain selection (context menu)
  Right click text on any page : “StudyPilot: Explain selection”.

• Export

  • Copy all summaries
  • Download .md
  • Download PDF (opens print dialog; choose “Save as PDF”)

• Theme & persistence
  Auto / Light / Dark with proper icon, plus saved preferences for theme and language.

• Local-first by default
  Uses Chrome’s built in AI models (on device) when available. No backend.

---

How it works (APIs)

• Summarizer API – key-point summaries in Markdown
• Prompt API (LanguageModel) – flashcards, Q&A, translation fallback, diagram explain
• Translator API – preferred path for translations
• Rewriter API – first choice for “Explain selection”, with Prompt fallback

The extension checks availability at runtime and gracefully falls back where possible.

---

Requirements

• A version of Google Chrome that supports the Built-in AI APIs (Early Preview / recent Chrome builds).
• On first use, Chrome may download the small on device model. If you see a tip like “Click ‘Summarize’ to allow model start”, click once and retry after the model initializes.

If the APIs are not available on your device/build, the panel will tell you. The extension never sends page content to a server.

---

Install (Developer Mode)

1. Clone or download this repository.
2. Open chrome://extensions and toggle Developer mode (top right).
3. Click Load unpacked and select the project folder (the one with manifest.json).
4. Pin StudyPilot and open any web article to use it.

---

Usage

1. Click the StudyPilot popup on any article page.
2. Pick a language and hit Summarize.
   • First run may trigger a model download; try again once it’s ready.
3. Explore the tabs:
   • Summary – sectioned notes (copy / open section / export)
   • Flashcards – “Generate flashcards” → download CSV
   • Ask – ask short questions about the page content
   • Diagram Q&A – upload an image and ask a question
   • Explain selection – paste text or use the right-click action

Export

• Copy all – one combined Markdown document
• Download .md – a Markdown file
• Download PDF – opens the print dialog (if blocked, press Ctrl/Cmd+P)

---

Files & Structure

• manifest.json # MV3 manifest
• panel.html/.css/.js # Main UI panel
• content.js # Receives “Explain selection” handoff
• sw.js # Service worker (context menus / messaging)
• ai.html / ai.js # Optional diagnostics / internal page
• icon.png # Extension icon
• PRIVACY.md # Full privacy policy
• LICENSE # MIT

---

Permissions

• activeTab – interact with the current page to extract sections
• scripting – inject a tiny snippet for section detection and “Go to section”
• storage – save theme/language preferences
• contextMenus – right click “Explain selection”
• host_permissions: ["http://*/*", "https://*/*"] – read text of the page you’re summarizing

Privacy: StudyPilot runs entirely in your browser. No remote server is used for AI processing. See PRIVACY.md.

---

Troubleshooting

• “Unsupported page” – works on http(s) pages. chrome://, edge://, PDFs, and some web apps may not expose readable text.
• Model unavailable / activation – click Summarize once to allow Chrome to initialize or download the on device model, then try again.
• PDF dialog didn’t show – some setups block programmatic print. Press Ctrl/Cmd+P.
• Dark print preview – your OS/browser print theme may default to dark. Switch to light before saving (known UX todo).

---

Roadmap

• Auto open print dialog more reliably across platforms
• Force light theme for printed/PDF output
• More languages + per-section copy/export
• Keyboard shortcuts
• Optional in panel screenshot to diagram flow

---

Built with

• Chrome Summarizer API
• Chrome Prompt API (LanguageModel; multimodal)
• Chrome Translator API
• Chrome Rewriter API
• Manifest V3 + Vanilla JS + zero backend

---

License

Released under the MIT License. See LICENSE.

---

Acknowledgements

Thanks to the Chrome team for the Built-in AI Early Preview and APIs that make private, local-first UX possible.