Databricks integration

Analytics & Data · 41 actions · Custom auth

Manage Databricks jobs, runs, SQL warehouses, and vector search indexes.

Put Databricks to work View Databricks integration docs

Overview

The ModuleX Databricks integration lets a ModuleX agent operate Databricks on your behalf — running create, update and cancel actions across all runs, runs and endpoints — directly from a plain-English request, using your organization's own Databricks credentials. No pre-built workflow is required: the agent picks the right Databricks action for the task.

Databricks is an analytics & data platform. ModuleX adds the agent layer: ask for an outcome and it selects and runs the right Databricks action — or, when you want a repeatable process, the composer assembles a Databricks workflow for you, streaming the nodes onto the canvas as it builds.

Ask the assistant

Drive Databricks in plain English

Type what you want. A ModuleX agent picks the right Databricks action — or chains several — and runs it. No workflow to build.

Create a new vector search endpoint
resolves tocreate_endpoint
Find the run I describe and cancel it
resolves toget_runcancel_run
Cancel all active runs for a job
resolves tocancel_all_runs
Show me the endpoints that match what I describe
resolves toget_endpoint

Use cases

What you can automate with Databricks

Find the right run and cancel it in one promptget_runcancel_run
Cancel an all run on requestcancel_all_runs
Create an endpoint straight from a requestcreate_endpoint
Remove a job you no longer needdelete_job
Edit a sql warehouse from a single promptedit_sql_warehouse
Look up a job permission on demandget_job_permissions

At a glance

Databricks integration at a glance

VendorDatabricks

CategoryAnalytics & Data

Actions available41

AuthenticationCustom

Uses your own credentialsYes

Works with the assistantYes

Works in the composerYes

Multi-step / tool-chainingYes

Technical referenceView docs

Integration version1.0.0

Last updatedJun 2026

Reference

All 41 Databricks actions

All Runs1

cancel_all_runs

Cancel all active runs for a job.

Runs5

cancel_run

Cancel a job run. The run is canceled asynchronously.

delete_run

Delete a non-active run. Returns an error if the run is active.

export_run

Export and retrieve the job run task.

get_run

Retrieve the metadata of a run.

list_runs

List all runs available to the user.

Endpoints4

create_endpoint

Create a new vector search endpoint.

delete_endpoint

Delete a vector search endpoint.

get_endpoint

Get details of a specific vector search endpoint.

list_endpoints

List all vector search endpoints.

Jobs5

create_job

Create a new Databricks job.

delete_job

Delete a job. Deleted jobs cannot be recovered.

get_job

Retrieve the details for a single job.

list_jobs

List all jobs using automatic pagination.

update_job

Update an existing job. Only the fields provided will be updated.

SQL Warehouses7

create_sql_warehouse

Create a new SQL Warehouse in Databricks.

delete_sql_warehouse

Delete a SQL Warehouse by ID.

edit_sql_warehouse

Edit the configuration of an existing SQL Warehouse.

get_sql_warehouse

Retrieve details for a specific SQL Warehouse.

list_sql_warehouses

List all SQL Warehouses available in the workspace.

start_sql_warehouse

Start a SQL Warehouse by ID.

stop_sql_warehouse

Stop a SQL Warehouse by ID.

Vector Search Indexes6

create_vector_search_index

Create a new vector search index in Databricks.

delete_vector_search_index

Delete a vector search index.

get_vector_search_index

Retrieve details about a specific vector search index.

list_vector_search_indexes

List all vector search indexes for a given endpoint.

query_vector_search_index

Query a specific vector search index.

sync_vector_search_index

Synchronize a Delta Sync vector search index.

Vector Search Index Datas2

delete_vector_search_index_data

Delete rows from a Direct Access vector index by primary-key values.

upsert_vector_search_index_data

Upsert data into an existing vector search index.

Job Permissions2

get_job_permissions

Get permissions of a job.

set_job_permissions

Set permissions on a job.

Run Outputs1

get_run_output

Retrieve the output and metadata of a single task run.

SQL Warehouse Configs2

get_sql_warehouse_config

Retrieve the global configuration for SQL Warehouses.

set_sql_warehouse_config

Update the global configuration for SQL Warehouses.

SQL Warehouse Permissions2

get_sql_warehouse_permissions

Retrieve the permissions for a specific SQL Warehouse.

set_sql_warehouse_permissions

Update the permissions for a specific SQL Warehouse.

Repair Runs1

repair_run

Re-run one or more tasks.

Reset Jobs1

reset_job

Overwrite all settings for a job.

Job Nows1

run_job_now

Run a job now and return the ID of the triggered run.

Scan Vector Search Indexes1

scan_vector_search_index

Scan a vector search index and return entries after a given primary key.

See full parameters and response schemas in the Databricks integration docs

Two ways

Two ways to use Databricks in ModuleX

Ask the assistantType what you want done and a ModuleX agent picks the right Databricks action and runs it — no workflow to build.

Compose a workflowNeed it to happen every time? Describe the process and the composer wires Databricks into a repeatable workflow you can run on a schedule, from chat, or as an API.

Security

Connecting Databricks

CustomDatabricks uses a custom authentication flow. ModuleX encrypts whatever credentials you supply and scopes them to your organization.

Step-by-step setup in the Databricks docs

Pairs well with Databricks

Agents often chain Databricks with these — connect them once and the agent can use all of them in a single task.

PostHog78 actions

Amplitude11 actions

Microsoft Power BI10 actions

Google Analytics6 actions

Google Tag Manager6 actions

Segment6 actions

FAQ

Databricks + ModuleX FAQ

A ModuleX agent can run any of Databricks's 41 actions — across all runs, runs, endpoints and jobs — from a plain-English request, using your organization's own Databricks credentials.

No. The agent selects the right Databricks action from your prompt. A workflow is only needed when you want a repeatable process — and the composer builds that for you from a description.

Through a secure Databricks connection; your credentials are stored encrypted and scoped to your organization.

41, spanning all runs, runs, endpoints and jobs.

Databricks runs on your own credentials, so any Databricks usage is billed by Databricks directly. ModuleX meters the agent run — and if you connect your own model keys, there's no ModuleX markup on model usage.

Put Databricks to work in ModuleX.

Connect Databricks once with your own credentials and let your agent run all 41 actions on demand.

Put Databricks to work View technical docs

Last updated: Jun 2026Browse all 175 integrations →