thucnc / pdf-to-json

GPT-3 & Next.js - JSONify a PDF - Convert PDFs into JSON data with your own custom schema.

Working with PDFs can be a huge drag. With AI, we can take PDFs and extract custom JSON data which make them much easier to work with.

Check out a live deployment of this app at jsonify.info.

Tech stack includes Next.js, Tailwind, OpenAI, and Typescript. Note that this app does not use LangChain, Pinecone, or even GPT-4 (it is using the GPT-3.5 model). While you would likely want to consider these on a more robust app, the purpose of this app is to simply demonstrate the "JSONification of a PDF" process.

Get in touch via twitter if you have questions

Development

1. Clone the repo

git clone https://github.com/AustonianTX/pdf-to-json.git

2. Install packages

npm install

3. Set up your .env file

• Create a .env file, and add your OpenAI API key

OPENAI_API_KEY=

• Visit openai to retrieve API keys and insert into your .env file.

Run the app

One you have installed the dependenies, you can run the app using npm run dev to launch the local dev environment. From there you can define a schema, upload your PDF file(s), and parse them into JSON.

Troubleshooting

In general, keep an eye out in the issues and discussions section of this repo for solutions.

Credit

The frontend of this repo was built with Austin Taylor at the Austin AI X Foundation Capital hackathon.