Spaces:
Configuration error
Configuration error
π PDF Chatbot β Runs in Your Browser
A zero-server, zero-API-key chatbot that lets you ask questions about your PDFs using lightweight AI models running entirely in your browser.
π Live Demo: https://huggingface.co/spaces/ghostofpk20/pdf-chatbot-browser
π Try It Now (No Installation!)
Just open the link above and start chatting with your PDFs. That's it.
First use only: The AI embedding model (~22MB) downloads from Hugging Face. After that, it works completely offline.
π§ How It Works
| Step | What Happens |
|---|---|
| 1. Upload PDF | PDF.js extracts text entirely in your browser |
| 2. Chunk & Index | Text is split into chunks; each chunk gets an AI embedding vector |
| 3. Ask Question | Your question is embedded; top-K matching chunks are retrieved |
| 4. Get Answer | Best sentences from matching chunks are extracted and returned |
All processing happens locally β your PDFs never leave your device.
π― Key Features
- β Fully offline after first download β no servers, no API keys
- β Drag & drop PDF upload
- β Multiple PDFs β chat with any uploaded document
- β WebGPU acceleration β 5-10x faster on Chrome/Edge with GPU
- β WASM fallback β works on any modern browser
- β Keyword fallback β still searches if AI model fails
- β Sources shown β every answer includes the document excerpts used
- β No scanned PDFs β text-based PDFs only (browser limitation)
π¦ Chrome Extension (Bonus)
Download chrome-extension.zip from the files tab and load it as an unpacked extension in chrome://extensions/.
π οΈ Models Used
| Model | Size | Purpose |
|---|---|---|
Xenova/all-MiniLM-L6-v2 |
~22MB (q8) | Text embeddings for document retrieval |
- ONNX quantized to 8-bit for minimal memory
- Runs via Transformers.js + ONNX Runtime Web
- Answers are extractive β no LLM hallucinations, just facts from your document
π Privacy
- Zero network calls after model download
- Your PDFs never leave your browser
- No analytics, no tracking, no API keys
- Everything stays in browser memory only
π Browser Support
| Browser | Status |
|---|---|
| Chrome 113+ | β Full (WebGPU) |
| Edge 113+ | β Full (WebGPU) |
| Firefox | β οΈ WASM fallback |
| Safari 17.4+ | β οΈ WebGPU in preview |
π Credits
- Transformers.js β ML in the browser
- PDF.js β PDF text extraction
- ONNX Runtime Web β Model inference