LangChain Changelog

Sign up for our newsletter to stay up to date

Oops! Something went wrong while submitting the form.

Powered by LaunchNotes

Loading...

March 2025

Structured output improvements for multiple models

A batch of improvements has been released for structured output integrations: The release covers three common methods for structured output:

Standardized forced tool calling in LangChain

New standards have been implemented for forcing tool calls in LangChain chat models. The implementation allows for: See the docs:...

Prompt engineering LangGraph Studio

You can now view and edit your node logic (such as prompts) directly within the Studio UI. See which nodes use specific assistant fields and edit them...

Trace with the OpenAI Agents SDK

You can now connect your OpenAI Agents to LangSmith with just 2 lines of code ! The OpenAI Agents SDK enables developers to build agentic applications...

Renaming an experiment in the LangSmith UI

You can now rename experiments in LangSmith, making it easier to keep your workspace organized. Whether you're running quick tests in the Playground or...

LangGraph BigTool for providing agents access to a large number of tools

langgraph-bigtool is a Python library for creating LangGraph agents that can access large numbers of tools. Key features include: • Scalable access to tools:...

LangGraph Swarm for building multi-agent systems

We now have a lightweight library for building swarm-style multi-agent systems with LangGraph. Key features include: The library is available via pip install...

LangGraph 0.3: Pre-built agents

LangGraph has always been about giving full control over your AI agents—no hidden prompts, no enforced architectures. But we also know that getting started...

Claude 3.7 Support in LangChain Python & JavaScript

Claude 3.7 support has been released for both JavaScript (via @langchain/anthropic package) and Python versions of LangChain. This lets LangChain users to...

MCP Adapters for LangChain and LangGraph

The LangChain MCP Adapters is a package that makes it easy to use Anthropic Model Context Protocol (MCP) tools with LangChain & LangGraph. It: Why use MCP...

Powered by LaunchNotes