LLMs are not perfect when calling tools. Models may attempt to call non-existent tools or fail to return parameters that match the requested schema. Strategies such as keeping the schema simple, reducing the number of tools passed at once, and using good names and descriptions can help mitigate this risk, but they are not foolproof. … Read more
LangGraph allows you to easily manage conversation memory in graphs. These operational guides demonstrate how to implement various strategies for this. Managing Conversation History One of the most common use cases for persistence is using it to track conversation history. It makes continuing conversations easier. However, as conversations get longer, this conversation history accumulates and … Read more
How to Handle a Large Number of Tools The subset of tools available for invocation is typically determined by the model (although many providers also allow users to specify or limit the selection of tools). As the number of available tools increases, you may want to limit the LLM’s selection to reduce token consumption and … Read more
LangGraph has built-in first-class stream support. Streaming output graph .stream and .astream are synchronous and asynchronous methods for streaming outputs back from the graph run. When calling these methods, several different modes can be specified (for example, `graph.stream(…, mode=”…”)`): values: This will transmit the full value of the state after each step of the graph. … Read more
Sometimes, we want the LLM that calls the tools to fill in a subset of parameters for the tool functions, while providing other values for the remaining parameters at runtime. If you are using LangChain-style tools, a simple way to handle this situation is to annotate the function parameters with InjectedArg. This annotation excludes the … Read more
How to Update Graph State via Tool A common use case is to update the graph state from within the tool. For example, in a customer support application, you may want to look up the customer account or ID at the start of the conversation. To update the graph state from the tool, you can … Read more
Andrew Ng recently said in his public lecture: I believe that AI Agent workflows will drive significant advances in artificial intelligence this year, possibly surpassing the next-generation foundational models. This is an important trend, and I urge everyone working in AI to pay attention to it. Coupled with the four paradigms for implementing Agent workflows … Read more
An AI travel assistant built on LangGraph: ai-travel-agent, which can search for flights, book hotels, and organize itineraries to email you. It has a memory capability, able to remember user interaction history and continue from where the last conversation left off. It adopts a human-machine collaboration model, allowing users to control key operations, such as … Read more
Introduction on how to use LangGraph to improve RAG. Long press to follow “Python Learning Base”, join the reader group, and share more wonderful content. 1. Introduction LangGraph is the latest member of the LangChain, LangServe, and LangSmith series, aimed at building generative AI applications using LLMs. Remember, all these are independent packages and must … Read more
Introduction Langchain recently launched an impressive course focused on LangGraph and its key features in developing powerful agent and multi-agent workflows. In this series, we will explore the essential insights from the course and create applications that leverage agent workflows. In the first part, we will cover the basic concepts of LangGraph and how to … Read more