LangChain: AI Framework

With the rapid advancement of AI, language models have become an integral part of modern applications. Even though fetching data via complex calculations done by APIs continues, the creative juice only flows when the LLMs kick in. LLMs allow us to build interesting applications backed by data that ideally would have taken us months to come up with.

But with each new technology comes newer challenges. Since models are being trained by different companies, each one is special and has its own set of capabilities that distinguishes it from other models. Some are open source while some are closed to the public, requiring a purchase.

However one of the biggest problems a developer faces is how to provide good prompts or inputs to the model to get the desired outcome. And once you have that outcome, how to build on that to get the next output. Something which the AI world calls — context.

That is where LangChain comes in. LangChain is a framework for building applications using LLMs and it provides you with all the necessary tools out of the box. E.g. when you are building a web application you might choose to go ahead with a framework like NestJs which has features already built in like routing, conventions etc. LangChain is similar but for AI applications.

The best part is that LangChain has Python and JavaScript packages already — hence catering to two of the biggest communities out there. The name is derived from the capability of the framework to allow the use of multiple models and pass the output of one model into another with ease thereby creating a chain.

Over the next couple of articles, we will be learning in-depth how the framework works and how it will make building AI-powered applications a breeze. So let's get started!