Apps & Databases

Databricks Apps (Python)

Skills: databricks-app-python MCP Tools: create_or_update_app, get_app, delete_app

Framework-Specific Apps

Create a Streamlit app on Databricks that connects to a SQL warehouse, queries
main.sales.transactions, and shows an interactive chart of revenue by month.

Build a Gradio app that provides a chat interface using the Databricks foundation
model API with the meta-llama/llama-4-maverick-17b-128e-instruct model.

Create a FastAPI REST service on Databricks Apps that exposes endpoints for CRUD
operations on a Lakebase PostgreSQL database.

Build a Dash app that visualizes real-time streaming data from a Unity Catalog
table with auto-refresh.

Create a Flask app with user authentication that lets users query their own data
through a form interface.

App Management

Deploy my app to Databricks Apps and show me the URL when it's live.

List all deployed Databricks Apps and show their status.

Update my existing app "analytics-portal" with the latest code changes.

Integration Patterns

Build a Streamlit app that:
1. Connects to a SQL warehouse for reading analytics data
2. Connects to Lakebase for storing user preferences
3. Uses a model serving endpoint for AI-powered recommendations

Create a Python app that uses Databricks OAuth (user-level auth) so each user
sees only the data they have access to in Unity Catalog.

---

Databricks Apps (AppKit / TypeScript)

Skills: databricks-agent-skill-databricks-apps MCP Tools: create_or_update_app, get_app

Scaffold a new AppKit application with analytics mode — I want to build a
dashboard that queries Unity Catalog tables.

Create an AppKit app in Lakebase mode with tRPC procedures for managing a task
list stored in PostgreSQL.

Build an AppKit app with SQL files that query main.sales.orders and
main.sales.customers, then generate the TypeScript types automatically.

Create an AppKit app with pre-built UI components showing a data table with
pagination, filters, and a chart.

---

Lakebase Autoscale (Managed PostgreSQL)

Skills: databricks-lakebase-autoscale, databricks-agent-skill-databricks-lakebase MCP Tools: create_or_update_lakebase_branch, delete_lakebase_branch, create_or_update_lakebase_sync, delete_lakebase_sync

Database Management

Create a new Lakebase Autoscaling project called "app-backend" with
scale-to-zero enabled and autoscaling from 0.5 to 8 CU.

Create a development branch of my Lakebase database for testing schema
migrations without affecting production.

Delete my dev branch "feature-xyz" after merging the schema changes to the
main branch.

Reverse ETL (Synced Tables)

Set up a reverse ETL sync from Delta table main.analytics.user_segments to my
Lakebase database so my app always has the latest segment data.

Remove the sync for the table "user_segments" from my Lakebase database.

Connectivity

Generate OAuth credentials for connecting to my Lakebase database from a Python
application.

Show me how to connect to my Lakebase Autoscale database from a Databricks App
using OAuth tokens.

Build a FastAPI app on Databricks that uses Lakebase Autoscale as its OLTP
backend with connection pooling.

---

Lakebase Provisioned (OLTP PostgreSQL)

Skills: databricks-lakebase-provisioned MCP Tools: create_or_update_lakebase_database, get_lakebase_database, delete_lakebase_database, generate_lakebase_credential

Create a new Lakebase Provisioned database with CU_2 capacity for my
application's transactional data.

List all Lakebase databases in my workspace and show their status and
capacity tier.

Generate connection credentials for my Lakebase database so I can connect from
my local development environment.

Set up reverse ETL from main.gold.product_catalog to my Lakebase Provisioned
database for serving low-latency lookups.

Build a Streamlit app that uses Lakebase Provisioned for storing user session
data and chat history for an AI agent.

Delete my Lakebase test database "dev-scratch" and all its resources.