LangGraph CLI Tutorial: Build an Agent in Minutes

Most developers skip the LangGraph CLI — and then spend weeks building their own API, streaming, memory and chat UI around their agent. In this video I'll show you how one config file and one command turn the ZEK AI shopping agent into a full Agent Server. We go from zero to a running server: installing the CLI, writing langgraph.json, plugging in a custom Postgres store with semantic search and a checkpointer for human-in-the-loop, running langgraph dev, and chatting with the agent in Agent Chat UI and LangSmith Studio — without writing a single line of frontend code. 💻 Source code (100% open source): https://github.com/Sami606713/zek 📚 LangGraph CLI docs: https://docs.langchain.com/langsmith/cli 📺 Backend episode:    • AI Agent Backend With FastAPI + LangGraph ...   ⏱️ Chapters 0:00 Welcome & recap of the ZEK series 0:38 Why most developers skip the LangGraph CLI 1:25 What you get: a full Agent Server 1:35 Installing the LangGraph CLI 1:47 The 5 commands: dev, build, deploy, dockerfile, up 2:48 Bonus: one-command project setup 4:58 langgraph.json explained (graphs, http, store, checkpointer, env) 5:59 Custom store: Postgres + semantic search with embeddings 7:10 Checkpointer: pause, ask a human, resume 8:07 Environment variables 8:49 Running langgraph dev 10:06 Agent Chat UI: chat with your agent, no custom UI 10:57 LangSmith Studio 11:47 Running the server with the MCP backend 12:29 Gotcha: authentication breaks the hosted UIs 12:58 The full Agent Server API at /docs 14:34 Built-in MCP endpoint 15:13 Recap & what's next 🧠 What you'll learn • What the LangGraph CLI is and why it matters for production agents • The 5 core commands and when to use each • How to configure langgraph.json for a real project • Adding a custom Postgres store with semantic search • Using a checkpointer for human-in-the-loop flows (like payments) • Testing your agent with Agent Chat UI and LangSmith Studio • Why auth breaks hosted UIs — and how your own frontend still works • The built-in Agent Server API and MCP endpoint ⏭️ Next: connecting the ZEK agent to the store frontend. 👍 If this saved you time, like, subscribe and share it with a developer building AI agents. Questions? Drop them in the comments! #LangGraph #LangChain #AIAgents #AgenticAI #Python #LangSmith #MCP #BuildInPublic #ZEK Follow along: Instagram:   / connectxeo   X: https://x.com/0xAgentHQ TikTok:   / connectxeo   YouTube:    / @connectxeo