Genie Code

What it does

Genie Code takes a natural-language request, plus context such as Unity Catalog tables, columns and lineage, notebooks, files, dashboards or uploaded screenshots. From these it plans a task, writes and runs code, and produces notebooks, SQL, Spark Declarative Pipelines, ML models and endpoints, and AI/BI dashboards. It uses cell outputs to refine results and fix errors.

People work with it inside the Databricks workspace: notebooks, the SQL editor, the Lakeflow Pipelines Editor, AI/BI dashboards and MLflow, or in a full-page command center with parallel chats. It builds a plan you can review before it executes. It differs from general coding agents by being grounded in Unity Catalog metadata and permissions. Teams can add custom instructions, reusable agent skills, cross-session memory and MCP servers.

It suits data engineers, data scientists and ML engineers already on Databricks; business users are pointed to Genie One or Genie Agents instead. Agentic features need partner-powered AI features enabled for the account and workspace, in a supported region; otherwise a Databricks-hosted model without agentic capabilities is used. Since July 8, 2026 it is billed pay-as-you-go with a per-user free monthly allowance. It sends prompts and data context to the model and is limited to what your permissions allow.

What it can do

• Automates exploratory data analysis and reporting in notebooks and SQL
• Builds and optimizes ETL and Spark Declarative Pipelines, and monitors and fixes them
• Engineers features, trains, evaluates and deploys ML models, and helps tune endpoints
• Plans and generates dashboards from data assets and metrics
• Reusable agent skills, custom instructions and memory across sessions
• Builds reviewable multistep plans before executing

Works with

• Unity Catalog
• MCP servers
• Databricks Notebooks
• Lakeflow Pipelines Editor
• AI/BI Dashboards
• MLflow

Example workflows

Exploratory analysis

Input: a question plus selected tables as context. Steps: Genie Code finds relevant data, plans the analysis, runs notebook and SQL steps and refines on outputs. Output: cleaned data and a shareable report.

Pipeline build

Input: a plain-language ETL description. Steps: review the agent plan, approve it, let it generate and run a Spark Declarative Pipeline. Output: a pipeline plus fixes for errors it detects.

Before you choose

• Prompts and context such as table names, data samples and cell outputs are sent to the model within your Unity Catalog permissions
• By default an AI classifier approves tool actions; you can switch to approving each action yourself
• Usage is billed; admins can manage budgets and cost controls

Listed under

• Data

Get started

Requires a Databricks workspace; trial available via Databricks

1. Use a Databricks workspace in a supported region and ask an admin to enable partner-powered AI features
2. Open Genie Code in a notebook, SQL editor or full-page view
3. Optionally add custom instructions, skills or MCP servers

Billed pay-as-you-go from July 8, 2026 with a per-user free monthly allowance; allowance size and rates are on the Genie pricing page.

Source: docs.databricks.com/aws/en/genie-code, checked 28 Oct 2026. Prices can change; check before you buy.

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.

• AI Agent for Data Teams | Databricks Genie Code ↗www.databricks.com
• Databricks Pricing: Flexible Plans for Data and AI Solutions | Databricks ↗www.databricks.com
• Genie Code | 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.