LangChain explained - The hottest new Python framework

if you're building AI apps with large language models right now there is no way around Lang chain link chain is currently one of the hottest new python Frameworks released just at the end of last year it already crossed 20 000 stars on GitHub and has a thriving community of contributors and creators who build new cool stuff with it every single day on top of that the creator of langchain just announced a 10 million dollar seat round so what is it that makes Lang chain so exciting so let's

take a look at Lang chain how it works and what you can do with it so Lang chain is a framework for developing end-to-end applications powered by large language models imagine you want to build an app on top of chat GTP or any other powerful language model and then you want to combine it with your own data say an ocean database PDFs or your emails and you also want to construct prompts based on the plain user input and then you also want to store the conversation history and maybe you want to combine the models with another model or give had access to Google search or Wikipedia to make it even more powerful this sounds like a lot to consider for an end-to-end application right well Lang chain makes all of this a whole lot easier it allows building applications with llms through composability and currently it provides six different key functionalities divided into different modules so let's take a look at some

short examples for each of those models first and foremost Lang chain provides a generic interface for many llms you can access models from open AI hugging face cohere and many more providers prompts this includes prompt management prompt optimization and prompt serialization you can for example Define prompt templates that take the user input and then create the final prompt for the model chains go beyond just one single llm call and are sequences of course in the simplest example you can for example chain together a prompt template and an llm but the possible combinations are almost endless here memory length chain provides a standard interface for memory and a collection of memory implementations for example you can easily store the message history of a chatbot indices this module contains many utility functions so that you can combine the model with your own Text data for example it provides document loaders to load the data from different sources like notion PDFs or emails and it provides Vector store interfaces to efficiently store the text and make it searchable and lastly agents and tools this is an extremely powerful module you can set up agents powered by large language models that can use tools like Google search Wikipedia or a calculator and if this is used correctly this can give your app unlimited Powers alright so this was a very quick overview of how this framework works if you want to see more Lang chain content then make sure to subscribe to our Channel because we post moral language and content soon so I hope to see you in the next video bye