Creating complex applications with LangGraph
Now we’ve made a basic translation application, where a user provides an answer to a templated prompt and the LLM provides a translation. For our next example, we’re going to build on this framework in a few key ways by designing a question-answering application that chains together several important capabilities:
- We will enable open-ended dialogue through a chatbot
- We’ll use a vector database to retrieve relevant documents to our query from an internal store
- We’ll add a memory that allows the bot to keep track of its interactions with us
- We’ll provide the ability for feedback from a human-in-the-loop user
- We’ll provide the ability to look on the internet for additional content in response to prompts
By doing so, we’ll move from specifying a chain, where commands are processed in a linear order, to graphs where LLM outputs are used to determine...
