
Already know some of this? A 2-minute test skips the lessons you know.
Take the placement testA free 12-lesson course built from freeCodeCamp.org's 3h 10m of YouTube video. 2 days at 90 minutes a day, with study notes and a quiz in every lesson. A 2-minute placement test skips what you already know.
Taught by freeCodeCamp.org on YouTube. Videos play in YouTube's own player, so the creator keeps every view.
Understand the purpose and syntax of TypedDict, Union, Optional, Any, and lambda functions to ensure type safety in LangGraph state definitions.
Identify the fundamental building blocks of LangGraph—state, nodes, graph, edges (including conditional edges), start/end points, tools, tool nodes, StateGraph, runnables, and message types—to conceptualize graph-based workflows.
Create a minimal LangGraph with a typed state schema, a greeting node that updates state via a reducer, start/end points, and compile/invoke the graph to see data flow through a single node.
Extend the state schema to accept multiple numbers and an operation, create processing nodes for list data, and invoke the graph with structured inputs.
Build a graph with three nodes (greeting, age, skills) connected by two add_edge calls to solidify understanding of sequential graph construction.
Implement a conditional graph using a router node and add_conditional_edges to route execution to addition or subtraction nodes based on the operation field in state.
Create a looping graph where a router node decides whether to continue processing (add/subtract) or end, using a lambda pass-through function and conditional edges to form a cycle back to the router.
Integrate an LLM (ChatOpenAI) into a LangGraph node, invoke it with a human message, and run a simple REPL loop; observe that the agent lacks conversation memory.
Use Union[HumanMessage, AIMessage] and the add_messages reducer to maintain full conversation history in state, enabling the agent to remember previous turns.
Build a ReAct-style agent that uses a router node with conditional edges to decide when to call tools (add, subtract, multiply) via a ToolNode, looping back to the agent until no tool calls remain.
Create a document drafting agent that uses update and save tools, a conditional edge loop, and a system prompt to collaborate with the user via text, optionally accepting pre-existing content.
Build a RAG system that loads a PDF, chunks it with overlap, embeds into a Chroma vector store, and uses a retriever tool within a two-agent graph (LLM agent + retriever agent) with conditional edges to answer grounded questions.