PDF Assistant

A powerful Chrome extension that replaces the default PDF viewer with an interactive interface, enabling on-device AI capabilities like summarization, table extraction, and multimodal chat directly within your browser. All processing happens locally for maximum privacy and speed.

🌟 Key Features

• Seamless PDF Integration: Automatically opens any PDF in a feature-rich, interactive viewer.
• On-Device Summarization: Select any text to generate key points, a TL;DR, or a headline summary.
• Visual Table Extraction: Simply draw a box around a table to extract its data into copyable JSON or CSV.
• Multimodal AI Chat: Combine selected text and an image area (like a chart) to ask complex, context-aware questions.
• In-Place Content Modification: Replace selected text with AI-generated content and download a new version of the PDF.
• Whole Document Analysis: Generate a comprehensive summary of the entire PDF document with a single click.
• 100% Private and Offline: All AI processing is done on your machine using Chrome's built-in models. Your documents and prompts never leave your computer.
• No API Keys Needed: Works out-of-the-box with zero configuration or account registration.

🎯 Why an On-Device PDF Assistant?

Unlike cloud-based AI tools that send your documents to external servers, this assistant is built on a privacy-first principle. By leveraging Chrome's built-in AI, it provides powerful features without compromising your data's security. It's designed for users who need intelligent document tools without the privacy trade-offs, API costs, or internet dependency.

⚡ Prerequisites

Before installing, please ensure your system meets the following requirements:

1. Google Chrome Version

   • Must be Chrome version ≥ 127

2. Chrome Flags Enabled

   • Navigate to chrome://flags and enable the following:
     • #prompt-api-for-gemini-nano
     • #prompt-api-for-gemini-nano-multimodal-input
     • #optimization-guide-on-device-model
   • Relaunch Chrome after enabling the flags for the changes to take effect.

🚀 Installation

This extension is not on the Chrome Web Store. To install it, you need to load it as an unpacked extension:

1. Clone the repository or download and unzip the project files:

   git clone https://github.com/Pravin-dev06/PDF-Assistant

2. Open Google Chrome and navigate to chrome://extensions/.
3. Enable "Developer mode" using the toggle in the top-right corner.
4. Click the "Load unpacked" button that appears on the top-left.
5. Select the root directory of the cloned project (the one containing manifest.json).
6. The extension will now appear in your list and is ready to use!

⚙️ How to Use

1. Summarize Selected Text

• Open a PDF in Chrome. The custom viewer will load automatically.
• Click and drag to select the text you want to summarize.
• In the "On-Device AI Assistant" modal that appears, choose your desired summary type and length.
• Click Generate. A new, modified PDF with the summary in place of the original text will be downloaded.

2. Extract a Table

• Click the Select Area button in the top toolbar.
• Draw a box precisely around the table you want to extract.
• The modal will default to the "Extract Table from Area" action. Click Generate.
• A new modal will appear, allowing you to copy the extracted data as JSON or CSV.

3. Ask a Question About a Chart (Multimodal)

• First, select text that provides context for the chart (e.g., its caption).
• When the modal appears, click the + Add Image Context button.
• Draw a box around the chart or image.
• Select the "Custom Prompt" action and type your question (e.g., "What was the trend between Q2 and Q4?").
• Click Generate. The AI's answer will replace the text you initially selected in the downloaded PDF.

4. Summarize the Entire Document

• In the top toolbar, click the Summarize Document button.
• A modal will show the progress. Once complete, a summary of the entire document will be displayed for you to copy.

💡 How It Works

The extension uses a modular architecture to provide a seamless experience:

1. content.js: A lightweight script that detects navigation to a PDF file and redirects the tab to the extension's custom viewer.html UI.
2. viewer.js: The core of the user interface. It uses PDF.js to render the PDF and handles user interactions like text and area selection. When a user requests an AI action, it calls the appropriate on-device Chrome AI API (Summarizer or LanguageModel).
3. background.js: The service worker manages PDF modification. When it receives AI-generated text from viewer.js, it uses the pdf-lib library to fetch the original PDF, erase the old content by drawing a white rectangle over it, and draw the new text in its place before triggering a download.

🔒 Privacy & Security

• 100% Offline Processing: All AI tasks run locally using Chrome's built-in Gemini Nano model.
• No External Servers: Your documents, selections, and prompts are never sent over the internet.
• No Tracking or Data Collection: Your activity remains private and on your device.

📝 License

MIT License

⚠️ Troubleshooting

If you encounter any issues:

1. Verify your Chrome version is 127 or newer.
2. Confirm that all three required flags are enabled in chrome://flags.
3. Ensure you have relaunched Chrome after enabling the flags.
4. Check that "Developer mode" is enabled on the chrome://extensions page.