Agent Bricks CLI (Beta): Build, Run Locally & Deploy a Databricks App with Memory & Sessions

Databricks has a new Agent Bricks CLI (Beta) 🚀 It lets you build a custom AI agent in code, run it on your laptop, and deploy it to Databricks Apps with real memory, saved sessions and MLflow traces. In this video I go from an empty folder to a live agent, step by step. I install the CLI with uv, scaffold a LangGraph agent with a chat app, add my own Python tool, test it locally with human approval and traces, and deploy it. I also show what happens on Databricks Free Edition, why deploy fails there, and how sessions and memory work on a trial workspace. ⚠️ Agent Bricks CLI is in Beta. On Databricks Free Edition, deploy fails with "Scale-to-zero must be enabled for this workspace tier", because the memory, session and runtime stores need Lakebase. Local dev works. For deploy, use a trial or paid workspace. ───────────────────────────────────── 🔥 WHAT YOU WILL LEARN ───────────────────────────────────── ✅ What the Agent Bricks CLI is and why you will not find it as a button in the Databricks UI ✅ How to install the CLI with uv ✅ How to log in with the Databricks CLI and agentbricks login ✅ How to scaffold a LangGraph agent with a ready-made chat app (agentbricks init) ✅ A tour of the project files, including agent.toml ✅ How to add your own Python tool (region_health.py) to the agent ✅ How to run and chat with the agent locally (agentbricks dev) ✅ How to pick another model when the default model is not available in your workspace ✅ MLflow traces of every model and tool call, logged automatically ✅ Human approval (human in the loop) before risky tool calls ✅ How to bind managed session and memory stores ✅ Free Edition limits: why deploy fails with the scale-to-zero error ✅ How to deploy to Databricks Apps on a trial workspace ✅ How sessions and memory are stored in Lakebase (Postgres) ✅ Sessions vs memory in action across two conversations ✅ CLI commands to list sessions and read memory entries ✅ How to stop the app and delete the session and memory stores ───────────────────────────────────── ⏱ CHAPTERS ───────────────────────────────────── 00:00 Agent Bricks Overview 01:56 Repo Setup Notes 02:49 Install CLI with UV 03:42 Databricks Auth Login 05:03 Agent Bricks Login 05:46 Scaffold LangGraph Agent 07:06 Project Files Tour 08:08 Add Custom Tool 11:25 Run Locally in Dev 14:06 Test Chat and Models 16:16 MLflow Traces and HITL 17:52 Bind Sessions and Memory 18:47 Deploy Free Tier Limits 20:15 Deploy on Trial Workspace 22:24 Apps and Lakebase Stores 27:07 Sessions Memory in Action 29:33 CLI Sessions Memory Commands 31:42 Stop App and Cleanup 33:34 Wrap Up and Next Steps ───────────────────────────────────── 🔗 RESOURCES ───────────────────────────────────── ↗ Agent Bricks CLI docs → https://docs.databricks.com/aws/en/agents/custom-agents/agent-bricks-cli ↗ Agent Bricks CLI on GitHub (full command reference) → https://github.com/databricks/databricks-ai-bridge/tree/main/integrations/agentbricks ↗ PyPI package → https://pypi.org/project/databricks-agentbricks/ ↗ Install uv → https://docs.astral.sh/uv/ ↗ Install the Databricks CLI → https://docs.databricks.com/aws/en/dev-tools/cli/install ↗ Databricks CLI authentication → https://docs.databricks.com/aws/en/dev-tools/cli/authentication ↗ Managed agent memory → https://docs.databricks.com/aws/en/agents/agent-memory/managed-memory ↗ Managed agent sessions → https://docs.databricks.com/aws/en/agents/agent-memory/managed-sessions ↗ MLflow tracing → https://docs.databricks.com/aws/en/mlflow3/genai/tracing/overview ↗ Free Edition limitations → https://docs.databricks.com/aws/en/getting-started/free-edition-limitations ↗ Demo code and files → https://github.com/sudarshan-koirala/agent-bricks-cli-demo --------------------------------------------------------------------- ☕ Buy me a Coffee: https://ko-fi.com/datasciencebasics ✌️Patreon: https://www.patreon.com/datasciencebasics ------------------------------------------------------------------------------------------ 🤝 Connect with me: 📺 Youtube: https://www.youtube.com/@datasciencebasics?sub_confirmation=1 👔 LinkedIn: https://www.linkedin.com/in/sudarshan-koirala/ 🐦 Twitter: https://twitter.com/mesudarshan 🔉Medium: https://medium.com/@sudarshan-koirala 💼 Consulting: https://topmate.io/sudarshan_koirala #databricks #agentbricks #aiagents #langgraph #mlflow #lakebase #databricksapps #datasciencebasics