Help us improve
Share bugs, ideas, or general feedback.
Create, manage, export, import, and query Databricks Genie Spaces for natural language SQL exploration on Unity Catalog data. Use for building spaces, migrations between workspaces, or Conversation API questions.
npx claudepluginhub databricks-solutions/ai-dev-kit --plugin databricks-ai-dev-kitHow this skill is triggered — by the user, by Claude, or both
Slash command
/databricks-ai-dev-kit:databricks-genieThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Create, manage, and query Databricks Genie Spaces - natural language interfaces for SQL-based data exploration.
Create and manage Databricks Agent Bricks: Knowledge Assistants for document Q&A via RAG, Genie Spaces for natural language to SQL, and Supervisor Agents for multi-agent orchestration. Use for conversational AI apps on Databricks.
Builds apps on Databricks Apps platform for dashboards, data apps, analytics tools, and visualizations. Evaluates analytics vs Lakebase data access patterns before scaffolding.
Analyzes lakehouse data interactively using Fabric Lakehouse Livy API sessions and PySpark/Spark SQL for DataFrames, joins, Delta time-travel, and JSON analysis.
Share bugs, ideas, or general feedback.
Create, manage, and query Databricks Genie Spaces - natural language interfaces for SQL-based data exploration.
Genie Spaces allow users to ask natural language questions about structured data in Unity Catalog. The system translates questions into SQL queries, executes them on a SQL warehouse, and presents results conversationally.
Use this skill when:
| Tool | Purpose |
|---|---|
manage_genie | Create, get, list, delete, export, and import Genie Spaces |
ask_genie | Ask natural language questions to a Genie Space |
get_table_stats_and_schema | Inspect table schemas before creating a space |
execute_sql | Test SQL queries directly |
| Action | Description | Required Params |
|---|---|---|
create_or_update | Idempotent create/update a space | display_name, table_identifiers (or serialized_space) |
get | Get space details | space_id |
list | List all spaces | (none) |
delete | Delete a space | space_id |
export | Export space config for migration/backup | space_id |
import | Import space from serialized config | warehouse_id, serialized_space |
Example tool calls:
# MCP Tool: manage_genie
# Create a new space
manage_genie(
action="create_or_update",
display_name="Sales Analytics",
table_identifiers=["catalog.schema.customers", "catalog.schema.orders"],
description="Explore sales data with natural language",
sample_questions=["What were total sales last month?"]
)
# MCP Tool: manage_genie
# Get space details with full config
manage_genie(action="get", space_id="space_123", include_serialized_space=True)
# MCP Tool: manage_genie
# List all spaces
manage_genie(action="list")
# MCP Tool: manage_genie
# Export for migration
exported = manage_genie(action="export", space_id="space_123")
# MCP Tool: manage_genie
# Import to new workspace
manage_genie(
action="import",
warehouse_id="warehouse_456",
serialized_space=exported["serialized_space"],
title="Sales Analytics (Prod)"
)
Ask natural language questions to a Genie Space. Pass conversation_id for follow-up questions.
# MCP Tool: ask_genie
# Start a new conversation
result = ask_genie(
space_id="space_123",
question="What were total sales last month?"
)
# Returns: {question, conversation_id, message_id, status, sql, columns, data, row_count}
# MCP Tool: ask_genie
# Follow-up question in same conversation
result = ask_genie(
space_id="space_123",
question="Break that down by region",
conversation_id=result["conversation_id"]
)
Before creating a Genie Space, understand your data:
# MCP Tool: get_table_stats_and_schema
get_table_stats_and_schema(
catalog="my_catalog",
schema="sales",
table_stat_level="SIMPLE"
)
# MCP Tool: manage_genie
manage_genie(
action="create_or_update",
display_name="Sales Analytics",
table_identifiers=[
"my_catalog.sales.customers",
"my_catalog.sales.orders"
],
description="Explore sales data with natural language",
sample_questions=[
"What were total sales last month?",
"Who are our top 10 customers?"
]
)
# MCP Tool: ask_genie
ask_genie(
space_id="your_space_id",
question="What were total sales last month?"
)
# Returns: SQL, columns, data, row_count
Export a space (preserves all tables, instructions, SQL examples, and layout):
# MCP Tool: manage_genie
exported = manage_genie(action="export", space_id="your_space_id")
# exported["serialized_space"] contains the full config
Clone to a new space (same catalog):
# MCP Tool: manage_genie
manage_genie(
action="import",
warehouse_id=exported["warehouse_id"],
serialized_space=exported["serialized_space"],
title=exported["title"], # override title; omit to keep original
description=exported["description"],
)
Cross-workspace migration: Each MCP server is workspace-scoped. Configure one server entry per workspace profile in your IDE's MCP config, then
manage_genie(action="export")from the source server andmanage_genie(action="import")via the target server. See spaces.md §Migration for the full workflow.
Before creating a Genie Space:
Use these skills in sequence:
databricks-synthetic-data-gen - Generate raw parquet filesdatabricks-spark-declarative-pipelines - Create bronze/silver/gold tablesSee spaces.md §Troubleshooting for a full list of issues and solutions.