LangGraph Made Easy

Create your first LangGraph line graph by wiring start to end through nodes and edges, defining a state container, and visualizing the hello world flow from input to output.

Convert a wiring graph into a living llm call by a single node that reads a prompt, invokes a chat model, and stores the response in state via graph invoke.

Understand conditional routing by building a runnable graph with four nodes and a router that decides the next hop at runtime, creating multiple endings and replayable variability.

Define a line graph state schema using type dict, data class, and pydantic to ensure runtime validation, clarity, and scalable data contracts.

Explore building a conditional graph with a type dict state for email and plan (premium or VIP), wire an upgrade router, and add runtime validations with guardrails to fail fast.

Explore how a data class state enables cleaner dot notation and contract flow, while Pydantic enforces runtime validation and upfront guards for email and plan in a conditional graph.

Explore how line graph handles state updates, collisions, and safe merging with reducers in LangGraph. See parallel branches, annotated lists, and custom reducers merge tallies without crashes.

demonstrate that two graph implementations - annotated list with an add message reducer and the built in messages state - produce identical, appended chat histories.

Turn a simple prompt into a runnable graph and scale to a travel planner using a typed trip state with city, days, budget, and an itinerary answer.

Learn how to implement and preserve multi-turn conversation history in a graph-based chat model using message state, the add message reducer, and a memory saver with checkpoints.

Discover how to empower an llm with tools, define and bind tools, and implement one-way and looping tool calls that translate model plans into safe, real-world actions.

Build and bind a text stats tool that analyzes a paragraph for word and sentence counts, reading time, top keywords, and enables the LLM to summarize results.

Build a staged travel agent by wiring an LLM to subway directions and attractions tools in a simple graph, enabling automatic tool calls to plan routes and surface attractions.

Surgically edit chat history by replacing messages with the same ID and removing them via remove message, using the add_messages reducer to maintain a tidy, duplicate-free timeline.

This lecture demonstrates token-budgeted chat by trimming history to a 120-token window, preserving only the most recent full messages before calling the model to generate concise, on-budget replies.

Create a memory aware chat loop that summarizes itself, keeps recent messages short, and passes conversation state across runs with a rolling memo.

Explore how an ai workflow separates public inputs and outputs from private scratch state, using plan and compose to transform and distill a rich state into a clean trip IO.

Explore a single-state graph workflow that carries all notes from brainstorm to final answer, using one schema frame and two refinements for quick prototyping and polished output.

Validate inputs and outputs with pydantic models, manage a rich internal trip workspace for deep work state, and publish a tidy itinerary from a graph with clean contracts.

Learn to build a human-in-the-loop graph that sketches an itinerary, pauses for review via interrupt, and resumes to publish a validated final plan with optional edits.

Persist state to SQLite with a check pointer to enable session memory, history preservation, and state inspection across restarts.

LangGraph Made Easy trains cross-thread memory using a short-term memory saver for per-thread chats and a long-term in-memory store that preserves profiles and recent trips in an evolving memo.

Learn how to extract structured, validated JSON from natural language using a simple schema, then store and leverage it in a long-term memory graph to personalize conversations.

Learn how trust call uses a line-graph based library to deliver reliable, validated structured extraction, patching and safely updating Pydantic schemas in evolving data for travel agents.

The course demonstrates building a memory-enabled graph from a simple line graph to a structured memory system, using SQLite check pointer and in-memory storage for personal, persistent chats.

Build a smart New York City travel concierge that turns plane requests into personalized, structured itineraries with day-by-day plans, dining and transit tips, blending automation with human review for accuracy.

Leverage a parallel AI orchestration graph and the send API to generate three trip ideas and complete itineraries for New York City, then merge results into a scalable, structured output.