Agent Bricks

What it does

Agent Bricks takes in enterprise data, schemas, business definitions and tool connections, and produces deployed AI agents that answer questions or take actions over governed data. Teams can build in an IDE with frameworks such as LangChain, LangGraph or LlamaIndex and deploy to serverless compute, or use managed no-code builders like Supervisor Agent and Knowledge Assistant. Agents can be served as REST APIs or run on scheduled workflows, with traces, evaluations and governance applied through Unity Catalog.

Teams use it to ground agents in lakehouse data, connect external systems through MCP, and route requests across models from OpenAI, Anthropic, Google, Meta and others, with fallbacks. Agents, models, MCP servers and tools are tracked and permissioned in a single governance layer, and each session can be traced and scored in MLflow, including with LLM-as-judge evaluation. Behavior and tool use can be validated in AI Playground before going live.

It suits data and platform teams on Databricks, from pro-code developers to business teams that need agents quickly. Access requires a Databricks account, and the free trial covers the full Data + AI Platform. Agent Bricks is in beta, and Databricks beta products are not intended for production use and carry no SLA, so confirm production readiness with a Databricks sales representative.

What it can do

• Build agents in an IDE with LangChain, LangGraph or LlamaIndex and deploy to serverless compute in one step.
• Use managed no-code agent builders such as Supervisor Agent and Knowledge Assistant for multistep reasoning over governed data.
• Connect agents to external systems through native MCP support, with credentials managed in Unity Catalog.
• Route requests across models from OpenAI, Anthropic, Google, Meta and others through one platform, with fallbacks.
• Trace, evaluate and monitor agent interactions with MLflow, including LLM-as-judge scoring.
• Govern models, tools, connections and agents with role-based access control and lineage in Unity Catalog.

Works with

• MCP (Model Context Protocol)
• LangChain
• LangGraph
• LlamaIndex
• SharePoint
• Google Drive
• Databricks Apps
• Lakebase

Example workflows

Governed data question agent (illustrative)

Input: a business question over lakehouse tables and documents. Steps: configure a Knowledge Assistant over governed data, validate tool use in AI Playground, then deploy as a REST API. Output: an agent that answers from permissioned data, with traces recorded in MLflow.

Developer agent with external tools (illustrative)

Input: a custom agent built in an IDE with LangGraph. Steps: connect external systems through MCP servers with credentials managed in Unity Catalog, deploy to serverless compute, and schedule runs on a workflow. Output: a deployed agent that calls approved APIs and SaaS tools with audit trails.

Before you choose

• Agent Bricks is in beta; beta products are not intended for production use and have no SLA, so confirm with a Databricks sales representative before relying on it in production.
• Pricing is usage-based, billed per second at the rates of the underlying Databricks products you use, so costs depend on the products and models involved. Request a quote for specific rates.
• The listed model providers may change over time, so check current model availability with Databricks.

Listed under

• Coding

Get started

The trial is for the full platform; the page does not state its length. Agent Bricks docs are linked for setup.

1. Create or sign in to a Databricks account and start the free trial of the Data + AI Platform.
2. Read the Agent Bricks docs and install the Agent Bricks CLI to scaffold an agent project.
3. Validate agent behavior and tool use in AI Playground, then deploy to serverless compute.

Costs depend on the underlying Databricks services, compute, cloud and region. Use the product’s current SKU rates and calculator; general platform trial credits are temporary.

We couldn’t find published prices for Agent Bricks. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 28 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Agent Bricks: build and govern production AI agents ↗www.databricks.com
• Beta Products | Databricks ↗www.databricks.com
• Develop agents on Databricks | Databricks on AWS ↗docs.databricks.com

Company website

databricks.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Web search

Confirms the official site

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 28 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.