langchain-ai/deepagents

Deep Agents is a ready-to-use AI agent framework built by the LangChain team. An AI agent is a program that uses a language model to plan and carry out multi-step tasks, not just answering a single question, but breaking down a goal, taking actions (like reading files, running commands, searching the web), and working toward a result. The project solves the problem of setup friction: instead of building all the plumbing yourself, connecting a language model to tools, managing conversation history, handling long context windows, Deep Agents gives you a fully wired agent out of the box with one function call. It comes pre-equipped with planning tools (to-do tracking), file system access (read, write, edit files), shell command execution, and the ability to delegate subtasks to separate sub-agents. You would use Deep Agents when building an automated assistant, a coding helper, or any system where you want an AI to complete complex multi-step tasks with access to tools and the file system. It also ships with a command-line interface that works similarly to AI coding tools, powered by your choice of language model. The framework is written in Python (with a TypeScript version available separately), built on LangGraph, works with any language model that supports tool use, and is licensed under MIT.