Introduction to LangChain - Python (LangChain Academy)
by LangChain
LangChain's own free course on building agents with create_agent, MCP tools, middleware and human-in-the-loop, with three projects to run in Jupyter.
Overview
Introduction to LangChain - Python is the foundation course in LangChain Academy, the training site run by LangChain, the company behind the LangChain and LangGraph frameworks. The course page lists about 1.5 hours of video over 30 lessons, and each module ends in a project you run from the companion GitHub repository, langchain-ai/lca-lc-foundations. Module 1, Create Agent, covers models, tools, short-term memory and multimodal messages, and its project is a personal chef agent. Module 2, Advanced Agent, adds the Model Context Protocol, context and state management, and multi-agent systems, with a wedding planner project; its first lesson starts a local MCP server with uv. Module 3, Production-Ready Agent, covers middleware, managing long conversations, human-in-the-loop review and dynamic agents, ending in an email assistant project and an optional lesson on building an agent chat UI. LangSmith is used for tracing and debugging throughout. The notebooks need Python 3.12 or 3.13, an OpenAI key (the course mostly calls gpt-5-nano) and a Tavily key for search; Anthropic and Google keys are optional and appear only in the first lesson. The repository went live on December 19, 2025, and its maintainers were still updating model versions and pinning MCP dependencies in July and August 2026.
At a Glance
- Topic
- Frameworks
- Level
- Beginner
- Format
- Course
- Cost
- Free
- Duration
- ~1.5 hours of video across 30 lessons, plus three notebook projects; self-paced
- Provider
- LangChain
- Hands-on
- Yes — code/exercises
- Certificate
- None
What You’ll Learn
- ✓Build a tool-calling agent with LangChain's create_agent API and short-term memory
- ✓Pass multimodal messages and swap model providers such as OpenAI, Anthropic and Google
- ✓Connect an agent to tools exposed by a local Model Context Protocol server
- ✓Manage agent context and state, and split work across a multi-agent system
- ✓Add middleware to an agent and keep long conversations within the context window
- ✓Insert human-in-the-loop approval steps and build dynamic agents for an email assistant
- ✓Trace and debug agent runs with LangSmith Observability and LangGraph Studio
Highlights
- •Written by the LangChain team against the current create_agent and middleware API, so the code matches today's docs rather than the pre-1.0 chains most older tutorials use
- •Every module ends in a runnable project (personal chef, wedding planner, email assistant) instead of isolated snippets
- •Companion repo langchain-ai/lca-lc-foundations has about 680 stars and 520 forks, with fixes merged as recently as August 2026
- •Covers MCP and human-in-the-loop inside a beginner course, topics many intro courses skip
- •Free to enroll; the main running cost is OpenAI API calls on a small model
Who It’s For
Best For
- ✓Python developers building their first LLM agent who want the framework's official path
- ✓Engineers migrating older LangChain chains or AgentExecutor code to the create_agent API
- ✓Teams standardizing on LangChain and LangSmith who need a shared onboarding course
- ✓Learners preparing for the LangChain Certified Agent Engineer prerequisites
Prerequisites
- •Working Python, including virtual environments and Jupyter notebooks
- •An OpenAI API key and a free Tavily API key (Anthropic, Google and LangSmith keys optional)
- •Python 3.12 or 3.13 and uv installed, needed for the MCP server lesson
FAQ
What is Introduction to LangChain - Python (LangChain Academy)?
Introduction to LangChain - Python is LangChain Academy's free foundation course for Python developers who want to build AI agents on the current LangChain agent API. Across three modules you build a personal chef agent, a multi-agent wedding planner that calls an MCP server, and an email assistant with middleware and human approval steps.
Is Introduction to LangChain - Python (LangChain Academy) free?
Introduction to LangChain - Python (LangChain Academy) is free to access.
What level is Introduction to LangChain - Python (LangChain Academy) for?
Introduction to LangChain - Python (LangChain Academy) is aimed at a beginner audience. Recommended background: Working Python, including virtual environments and Jupyter notebooks, An OpenAI API key and a free Tavily API key (Anthropic, Google and LangSmith keys optional), Python 3.12 or 3.13 and uv installed, needed for the MCP server lesson.
How long does Introduction to LangChain - Python (LangChain Academy) take?
Expect roughly ~1.5 hours of video across 30 lessons, plus three notebook projects; self-paced. Most learners work through it at their own pace.
What will I learn from Introduction to LangChain - Python (LangChain Academy)?
You'll learn: Build a tool-calling agent with LangChain's create_agent API and short-term memory; Pass multimodal messages and swap model providers such as OpenAI, Anthropic and Google; Connect an agent to tools exposed by a local Model Context Protocol server; Manage agent context and state, and split work across a multi-agent system; Add middleware to an agent and keep long conversations within the context window; Insert human-in-the-loop approval steps and build dynamic agents for an email assistant; Trace and debug agent runs with LangSmith Observability and LangGraph Studio.
Topics
Sources
This page was written from 5 sources, 3 on domains other than academy.langchain.com.