Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Answers to common questions about Agent Bricks, the Azure Databricks agent developer platform. For an overview of the platform, see What is Agent Bricks?.
What is Agent Bricks?
Agent Bricks is the Azure Databricks agent developer platform. It provides the infrastructure that custom agents need in production: governed access to models, MCP servers, and skills through Unity Gateway, an agent server and Agent Runtime to deploy and serve agents, managed memory and sessions, tools and MCP servers, MLflow tracing and evaluation, and governance through Unity Catalog.
Who is Agent Bricks for?
Agent Bricks is for developers who build custom agents in code, from a single agent to many agents that serve production traffic. To build agents without writing code, use Genie Agents.
How is Agent Bricks different from Genie Agents?
Genie Agents is a low-code way to build agents that answer questions over your tables and documents. Agent Bricks is a platform for writing and deploying your own agent code. A custom agent built on Agent Bricks can call Genie Agents as tools.
How is Agent Bricks different from Genie One?
Genie One is the experience for business users to work with data and AI without code. Agent Bricks is for developers who build custom agents. A custom agent can call Genie One as a tool.
How is Agent Bricks different from Omnigent?
Omnigent is a meta-harness that composes and coordinates multiple agents, such as coding agents. Agent Bricks provides the infrastructure to deploy and run custom agents. You can use Omnigent to develop custom agents that run on Agent Bricks.
What happened to Knowledge Assistant and Supervisor Agent?
Knowledge Assistant and Supervisor Agent are legacy offerings. For new low-code agents, use Genie Agents. For agents written in code, use the Agent Bricks CLI. See Legacy agent offerings.
Which frameworks and models does Agent Bricks support?
Agent Bricks works with any framework or harness. The Agent Bricks CLI includes templates for LangGraph and the OpenAI Agents SDK. Agents can use frontier and open models through Unity Gateway, which lets you switch models without changing agent code.
Where do agents built with Agent Bricks run?
Agents run on Agent Runtime, which runs the agent server on Databricks Apps. The agent server, DurableAgentServer, serves the agent through the invocation API and recovers runs that a crash or restart interrupts. See Deploy agents on Azure Databricks.
How do I get started with Agent Bricks?
Install the Agent Bricks CLI with pip install databricks-agentbricks. Then run agentbricks init to create a project, agentbricks dev to run the agent locally, and agentbricks deploy to deploy it. See Agent Bricks CLI.
Can I use Agent Bricks with an agent that I already built?
Yes. For agents built with LangGraph or the OpenAI Agents SDK, run agentbricks init --existing in the project directory. The CLI prepares migration instructions that a coding agent follows to convert the project. See Bring an existing agent.
How do I query an agent built with Agent Bricks?
Send requests to the invocation API at /api/invocations, with the Agent Bricks CLI, the REST API, or Python. See Query agents deployed on Azure Databricks.
How does Agent Bricks govern what agents can do?
Unity Catalog governs the data, tools, models, and agents that your agents use. Unity Gateway governs access to models, MCP servers, and skills, and applies guardrails, rate limits, and usage tracking to the calls that agents make to them. Tools can run with the permissions of the app's service principal or of the user who sent the request. See Request-user authorization.