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

Use the new GPT-4 api to build a chatGPT chatbot for Large PDF docs (56 pages used in this example).

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

Tutorial video

Get in touch via twitter if you have questions

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.

Development

Clone the repo

git clone [github https url]

Install packages

pnpm install

Set up your .env file

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

OPENAI_API_KEY=

PINECONE_API_KEY=
PINECONE_ENVIRONMENT=

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

In config/pinecone.ts, replace the PINECONE_INDEX_NAME with the index name you created on Pinecone. Edit the TOPICS to match your desired topics, namespaces, and prompts. There should be a matching folder in docs for each of your topics. THE FOLDER MUST EXACTLY MATCH THE TOPIC'S NAMESPACE AND MUST ONLY CONTAIN LOWER CASE LETTERS A-Z AND HYPHENS.
In utils/makechain.ts chain adjust the QA_PROMPT according to your use case. Change modelName in new OpenAIChat to a different api model if you don't have access to gpt-4. See the OpenAI docs for a list of supported modelNames. For example you could use gpt-3.5-turbo if you do not have access to gpt-4, yet.

Convert your PDF to embeddings

In docs add PDFs into the relevant folder for each topic. ChatGPT will show which PDF it referred to in the sources so give your PDFs descriptive names.
Run the script pnpm run ingest to 'ingest' and embed your docs
Check Pinecone dashboard to verify your namespace and vectors have been added.

Run the app

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

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
Make sure you're using the same versions of LangChain and Pinecone as this repo.
Check that you've created an .env file that contains your valid (and working) API keys.
If you change modelName in OpenAIChat note that the correct name of the alternative model is gpt-3.5-turbo
Pinecone indexes of users on the Starter(free) plan are deleted after 7 days of inactivity. To prevent this, send an API request to Pinecone to reset the counter.

Pinecone errors

Make sure your pinecone dashboard environment and index matches the one in your config folder.
Check that you've set the vector dimensions to 1536.
Switch your Environment in pinecone to us-east1-gcp if the other environment is causing issues.

Deleting a namespace

If you need to clear a namespace and to re-ingest it again you can edit the target namespace and the target index in scripts/delete-namespace.ts and then perform pnpm run delete-namespace

If you're stuck after trying all these steps, delete node_modules, restart your computer, then pnpm install again.

Credit