How to Get Started with LangChain

Photo by Austin Chan on Unsplash

I wanted to understand LangChain’s scope and capabilities. So, I recently read through every page of LangChain’s documentation. I was surprised by how much I didn’t know even after using it for several months. I’d been following the pattern in the quickstart guide. But, I realized that was more complex than it needed to be.

In my opinion, the first page of documentation should cover the easiest to integrate examples of the library that cover 80% of use-cases.

This guide’s goal is to provide that “first page” onboarding + some bonus info (consider this my attempt at a quickstart guide — v2).

Mission of LangChain

LangChain seeks to offer a single abstraction layer for every tool or library you may use in your effort to integrate with Large-Language Models.

For example, if you want to add a chatbot to your website that can answer questions that should exist somewhere in your documentation, you’re going to need a vector database. And even more than that, you’re going to need to:

1. Retrieve each page’s content
2. Embed each page’s content
3. Save the vector store
4. Load the vector store at runtime
5. Embed a user’s question