GPT-4 & LangChain - Create a ChatGPT Chatbot for Your PDF Files

Use the new GPT-4 api to build a chatGPT chatbot for multiple Large PDF files, docx, pptx, html, txt, csv.

Tech stack used includes LangChain, Chroma, Typescript, Openai, and Next.js. LangChain is a framework that makes it easier to build scalable AI/LLM apps and chatbots. Chroma is a vectorstore for storing embeddings and your PDF in text to later retrieve similar docs.

Tutorial video using the Pinecone db instead of the opensource Chroma db

The visual guide of this repo and tutorial is in the visual guide folder.

If you run into errors, please review the troubleshooting section further down this page.

Prelude: Please make sure you have already downloaded node on your system and the version is 18 or greater.

Features

• Support for multiple file formats: docx, pptx, html, txt, csv.
• Support for multiple files. Just put your files to docs folder, and run npm run ingest.
• Both GPT-3.5, GPT-4 are available. GPT-4 is recommended for better answer, while with slower response.
• Open source chromadb as vector database, you don't need to send your data to a cloud commercial vectordb.

Deployment

1. Clone the repo or download the ZIP

git clone [github https url]

2. Install packages

First run npm install yarn -g to install yarn globally (if you haven't already).

Then run:

yarn install

After installation, you should now see a node_modules folder.

3. Set up your .env file, and prepare chromadb server on your host.

• Copy .env.example into .env Your .env file should look like this:

OPENAI_API_KEY=
COLLECTION_NAME=

Note, the collection name must be an UUID which can be generated by uuid command in Linux/Mac.

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

• Visit chroma to run the chroma client locally using Docker.

git clone [email protected]:chroma-core/chroma.git
cd chroma
docker-compose up -d --build

4. In the .env file, replace the COLLECTION_NAME with a namespace where you'd like to store your embeddings on Chroma when you run npm run ingest. This namespace will later be used for queries and retrieval.

5. In utils/makechain.ts chain change the QA_PROMPT for your own usecase. Change modelName in new OpenAI to gpt-4, if you have access to gpt-4 api. Please verify outside this repo that you have access to gpt-4 api, otherwise the application will not work.

Convert your PDF files to embeddings

This repo can load multiple PDF files, and other files such as docx, pptx, txt, csv, html

1. Inside docs folder, add your pdf files or folders that contain pdf/docx/pptx files. There is an example legal case file in the docs folder already.

2. Run the script npm run ingest to 'ingest' and embed your docs. If you run into errors troubleshoot below.

3. Check the Chroma docker instance's dashboard log to verify a sucessful POST request has been made to the server.

Run the app

Once you've verified that the embeddings and content have been successfully added to your Chroma, you can run the app npm run dev to launch the local dev environment, and then type a question in the chat interface.

http://localhost:3000

Snapshot of the product

Just a few steps to deploy your server, prepare your files, and you can chat with your docs.

Troubleshooting

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

General errors

• Make sure you're running the latest Node version. Run node -v
• Try a different PDF or convert your PDF to text first. It's possible your PDF is corrupted, scanned, or requires OCR to convert to text.
• console.log the env variables and make sure they are exposed.
• Make sure you're using the same versions of LangChain and Chroma as this repo.
• Check that you've created an .env file that contains your valid (and working) API keys, environment and index name.
• If you change modelName in OpenAI, make sure you have access to the api for the appropriate model.
• Make sure you have enough OpenAI credits and a valid card on your billings account.
• Check that you don't have multiple OPENAPI keys in your global environment. If you do, the local env file from the project will be overwritten by systems env variable.
• Try to hard code your API keys into the process.env variables if there are still issues.

Credit

Frontend of this repo is inspired by gpt4-pdf-chatbot-langchain, and gpt4-pdf-chatbot-langchain-chroma