ChatPDF is a web app that can understand and answer questions about an uploaded PDF document. It serves as an assistant to help researchers, students, or professionals quickly find precise information from PDFs.
This is a Next.js project bootstrapped with create-next-app.
First, run the development server:
npm run dev
# or
yarn dev
# or
pnpm dev
# or
bun devOpen http://localhost:3000 with your browser to see the result.
- Upload a PDF and ask questions about its content
- Powered by AI to read and understand PDF documents
- Provides accurate answers to questions in plain language
- Built using Next.js, React, TypeScript, and other modern technologies
- Hosted on Vercel for fast, global access with minimal latency
- Leverages AWS S3, LangChain, Pinecone, and ChatGPT to understand documents and generate answers
- Improved UI/UX and logic flow to prevent Open Error on temp folder compared to original, plus user account management by adding ‘users’ table and making chat link ‘/chat/{userID}/{chatID}’ (instead of ‘/chat/{chatID}’)
A live demo is available here:
https://chatpdf-andrew.vercel.app/
Try it out by uploading a PDF or using one of the samples!
ChatPDF is currently hosted on Vercel and leverages its serverless platform for scale, speed, and ease of use. The front-end application is hosted on Vercel Edge locations around the world to minimize latency. PDF documents are stored in AWS S3 buckets in relevant regions.
The UI is a progressive web app that caches content like App Shell architecture. Pages render extremely quickly even for complex pages due to granular caching and static optimization.
1. Installation
2. Usage
3. Contributing
4. Credits
Click the upload button and select a PDF file from your computer. Wait for it to process (depending on size this takes 5-15 seconds).
Type questions related to the content of the PDF in plain English into the text prompt area - e.g. "What is the main topic of this research paper?". The AI will scan the uploaded document and attempt to return an accurate answer by analyzing the content.
Logged in users can access a personalized dashboard to manage conversations, PDFs uploaded, and questions asked. Share the dashboard link to keep discussions ongoing.
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
ChatPDF was inspired by Elliott Chong's YouTube Tutorial. It builds on his open-sourced project with UI and architectural improvements.
Huge thanks to:
- Elliot Chong
- Anthropic
- LangChain
- Pinecone
- OpenAI