Getting started with LangChain

As introduced in Chapter 2, LangChain is a lightweight framework meant to make it easier to integrate and orchestrate LLMs’ and their components within applications. It is mainly Python based, yet it recently extended its support to JavaScript and TypeScript.In addition to Large Language Models integration (which we will cover in the upcoming dedicated paragraph), we saw that LangChain offer the following main components:

• Models and prompt templates
• Data Connections
• Memory
• Chains
• Agents

Figure 1: LangChain’s components.

Next sections will be a deep dive for each of these components.

Models and Prompts

LangChain offers more than 50 integrations towards third-party vendor and platforms, including OpenAI, Azure OpenAI, Databricks and MosaicML, as well as the integration with Hugging Face Hub and the world of open-source LLMs. In Part 2 of this book, we will be trying various LLMs, both proprietary and open-source, leveraging those LangChain...