Custom Agents now available on Databricks

Agent Bricks Custom Agents, formerly Agent Framework, is now available on Databricks. Custom Agents allows developers to use their existing tools and workflows to rapidly build, test, and deploy production-quality AI agents as fully managed Databricks Apps. Developers don’t need to re-architect code or manage infrastructure. 

Teams can build agents locally, use AI coding tools, and iterate quickly on agent quality with tight feedback loops before deploying. Prebuilt agent skills, templates and integrated evaluation reduce setup work and help developers move from prototype to production faster.

Custom Agents integrate naturally into existing development workflows, including CI/CD pipelines, so teams can continuously test, refine, and ship improvements without re-architecting deployments. Developers can keep using their preferred models, frameworks, and tools, without rewriting code or adapting to vendor-specific environments.

When ready to deploy, agents run as Databricks Apps on serverless compute with built-in security and governance. This provides a consistent production environment without managing infrastructure or coordinating handoffs to platform teams.

Built-in memory powered by Lakebase gives developers a production-grade way to store and reuse agent state and conversation history, so agents stay context-aware across steps and sessions without extra infrastructure. Agents also connect directly to enterprise data and systems with consistent governance across data, models, and agent operations, reducing custom integration work and helping teams ship trusted, domain-aware agents faster.

Custom Agents is available now. Check out the documentation to get started.