This server acts as a Model Context Protocol (MCP) interface for Databricks, allowing LLM agents to manage and query a Databricks workspace through various APIs. Key capabilities include:
Authentication & Utilities
Identify the authenticated principal (whoami), view server configuration (auth_config)
Workspace Management
List, get, create directories, delete, export, and import notebooks and files; list/read files under /Workspace/files
Clusters
List, get, create, edit, start, restart, resize, terminate, delete clusters; view cluster events; list node types and Spark versions
Jobs
List, get, create, update, reset, delete jobs; trigger runs, list/get/cancel/repair/export runs; cancel all active runs
SQL Warehouses
List, get, create, edit, start, stop, delete SQL warehouses
SQL Execution & Queries
Execute SQL statements, poll status, fetch results, cancel statements; list/get/create/update/delete saved queries; view query history
Alerts & Dashboards
SQL Alerts: list, get, create, update, delete
Lakeview Dashboards: list, get, create, update, delete, publish, unpublish, restore, purge; manage schedules
Legacy Dashboards: list and get
Unity Catalog
CRUD operations on catalogs, schemas, tables, volumes, and functions
Additional Features
Delta Live Tables: list, get, create, start, stop, delete pipelines; manage pipeline updates
MLflow: manage experiments, runs, registered models, model versions, webhooks
Model Serving: create, list, get, update, delete, query serving endpoints
Vector Search: CRUD on endpoints and indexes; upsert, query, scan
Databricks Apps: list, get, create, update, delete
Repos (Git): list, get, create, update, delete; pull, push, commit
Secrets: manage scopes and secrets (list, put, get, delete)
DBFS: list, get, put, delete files
Tokens: list, create, revoke
Permissions (ACLs): get, set, update, delete on jobs, clusters, pipelines, etc.
Identity/SCIM: list users, groups, service principals
Delta Sharing: manage shares, recipients, providers
Genie (AI/BI): list spaces, ask questions
Account APIs: manage workspaces, metastores, storage, credentials, users, groups, service principals
Provides tools to manage and query Databricks workspaces, including clusters, jobs, SQL warehouses, notebooks, MLflow, Unity Catalog, Delta Live Tables, and more.
Supports dbt tasks in job creation and execution.
Enables management of Delta Live Tables (pipelines) and Delta Sharing shares, recipients, and providers.
Supports git-based repositories (Repos) with operations to list, get, create, update, delete, pull, push, and commit.
Provides tools for managing MLflow experiments, runs, models, registered models, model versions, and webhooks.
Allows management of Unity Catalog resources including catalogs, schemas, tables, columns, volumes, functions, and grants.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@databricks-mcp list my active clusters"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A Model Context Protocol (MCP) server that exposes the Databricks REST APIs as MCP tools so an LLM agent can manage and query a Databricks workspace.
The server speaks MCP over stdio (for Claude Desktop / Claude Code /
Cursor / IDE hosts) and supports a single --transport streamable-http mode
for remote deployments (Databricks Apps, container, etc.).
Safe by default. Out of the box only list and create tools are
exposed — roughly 110 of the 303 available. Mutating, destructive, and
single-record-read tools are hidden until you opt in via env vars or CLI
flags. See Permissions below.
303 tools across these Databricks domains. Default visibility indicates whether the tools are exposed in safe-by-default mode; everything else is opt-in.
Domain | Tools | Default |
Workspace | list / get / create / delete / export / import notebooks; list / get / mkdir / delete workspace files and dirs | list, create only |
Clusters | list, get, create, start, terminate, restart, resize, edit, delete; cluster events; cluster policies; instance pools; node types; spark versions | list, create only |
Jobs | list, get, create, run-now, list-runs, get-run, cancel-run, delete; full task types (notebook, spark_jar, python_wheel, dbt, sql, pipeline, run_job, condition, for_each) | list, create only |
SQL Warehouses | list, get, create, start, stop, edit, delete | list, create only |
SQL Queries / Dashboards / Alerts / Data | list / get / run SQL statements, dashboards (legacy + Lakeview), alerts | list, create only |
Unity Catalog | catalogs, schemas, tables, columns, volumes, functions, grants, model versions | list, create only |
Delta Live Tables (Pipelines) | list, get, create, start, stop, delete; pipeline updates | list, create only |
MLflow | experiments, runs, models, registered models, model versions, webhooks | list, create only |
Model Serving | serving endpoints (create, list, get, update, delete, query) | list, create only |
Vector Search | endpoints (create, list, get, delete), indexes (create, list, get, delete, upsert, query, scan) | list, create only |
Databricks Apps | list, get, create, update, delete | list, create only |
Repos (Git) | list, get, create, update, delete; pull, push, commit | list, create only |
Secrets | list, put, get, delete scopes and secrets | list, create only |
DBFS | list, get, put, delete files | list, create only |
Tokens | list, create, revoke | list, create only |
Permissions | get / set / update / delete ACLs on jobs, clusters, pipelines, etc. | list, create only |
Identity / SCIM | list users, groups, service principals | list, create only |
Delta Sharing | list / create / update / delete shares, recipients, providers | list, create only |
Genie (AI/BI) | list spaces, ask-question | list, create only |
Account APIs | workspaces, metastores, storage, credentials, users, groups, service principals | list, create only |
Utilities | workspace status, current user, whoami, auth_config | always exposed |
To expose everything (legacy behaviour), set
DATABRICKS_MCP_ENABLE_VERBS=list,get,create,update,delete.
Related MCP server: Databricks MCP Server
# From source (this repo)
uv tool install .
# Or pipx / pip
pipx install .
# or
pip install .
# Or run directly with uvx
uvx --from . databricks-mcpThe server needs three environment variables (or CLI flags):
Variable | Required | Example |
| yes |
|
| one of PAT or OAuth must be set |
|
| OAuth M2M alternative to PAT | — |
| only for account-level APIs |
|
You can also pass them as CLI flags: --host, --token.
Each tool is classified into one of five verb categories. The default
policy exposes only list and create; the rest are hidden.
Category | Default | Approx. tools | Examples |
| ON | ~50 |
|
| ON | ~30 |
|
| OFF | ~60 |
|
| OFF | ~100 |
|
| OFF | ~60 |
|
In increasing precedence:
Knob | Env var | CLI flag |
Verb categories ON by default |
|
|
Per-tool allow (always shown) |
|
|
Per-tool deny (always hidden) |
|
|
Resolution per tool: deny wins over allow wins over verb category.
# Legacy behaviour — full surface exposed
DATABRICKS_MCP_ENABLE_VERBS=list,get,create,update,delete databricks-mcp
# Default + opt in to running jobs and editing clusters
DATABRICKS_MCP_ENABLE_TOOLS=jobs_run_now,clusters_edit databricks-mcp
# Default + hide notebook delete even though delete is OFF anyway
DATABRICKS_MCP_DISABLE_TOOLS=workspace_delete databricks-mcp
# Read-only audit agent
DATABRICKS_MCP_ENABLE_VERBS=list,get databricks-mcp
# CLI flag form (same as the first env-var example)
databricks-mcp --enable-verbs list,get,create,update,deleteVerify the active policy at runtime via the auth_config tool — it
returns a policy block with enabled_verbs, allowed_tools,
denied_tools, and registered_tool_count. Full reference:
docs/PERMISSIONS.md.
Three ready-made profiles are in
examples/claude_desktop_config.json:
databricks — full surface (every verb enabled).
databricks-restricted — default list+create plus destructive-tool denylist.
databricks-readonly — list and get only.
Add the one you want to
~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"databricks": {
"command": "databricks-mcp",
"env": {
"DATABRICKS_HOST": "https://dbc-1234567890.cloud.databricks.com",
"DATABRICKS_TOKEN": "dapi..."
}
}
}
}claude mcp add databricks \
--transport stdio \
--env DATABRICKS_HOST=https://dbc-1234567890.cloud.databricks.com \
--env DATABRICKS_TOKEN=dapi... \
-- databricks-mcpTo opt back into the full surface:
claude mcp add databricks \
--transport stdio \
--env DATABRICKS_HOST=https://dbc-1234567890.cloud.databricks.com \
--env DATABRICKS_TOKEN=dapi... \
--env DATABRICKS_MCP_ENABLE_VERBS=list,get,create,update,delete \
-- databricks-mcp# stdio (default)
uv run databricks-mcp
# streamable HTTP (for remote deployment / Databricks Apps)
uv run databricks-mcp --transport streamable-http --host 0.0.0.0 --port 8000npx -y @modelcontextprotocol/inspector databricks-mcpSet DATABRICKS_HOST and DATABRICKS_TOKEN in the Inspector's env form
and click Connect. Browse the tool list — by default you'll see only
list/create tools. To see everything, add
DATABRICKS_MCP_ENABLE_VERBS=list,get,create,update,delete to the env.
Try the auth_config tool first to verify your setup; it reports the
active policy as JSON.
docs/PERMISSIONS.md — full permission reference,
verb classification, common configurations.
docs/AUTH.md — PAT vs OAuth M2M, on-behalf-of flow.
docs/TOOLS.md — every tool grouped by domain.
docs/EXAMPLES.md — sample invocations.
docs/DEVELOPING.md — how to add new tools.
MIT.
Create account-level credentials (e.g. for storage).
Inputs: body (required). Returns: result
Delete account-level credentials.
Inputs: credential_id (required). Returns: result
List account-level credentials.
Inputs: no named parameters. Returns: result
Create an account-level SCIM group.
Inputs: body (required). Returns: result
Delete an account-level SCIM group.
Inputs: id (required). Returns: result
Fetch an account-level SCIM group.
Inputs: id (required). Returns: result
List account-level SCIM groups.
Inputs: filter, startIndex, count. Returns: result
Create a Unity Catalog metastore.
Inputs: body (required). Returns: result
Delete a Unity Catalog metastore (set force to forcibly remove).
Inputs: metastore_id (required), force. Returns: result
Fetch a metastore.
Inputs: metastore_id (required). Returns: result
List Unity Catalog metastores under the account.
Inputs: no named parameters. Returns: result
Patch a metastore.
Inputs: metastore_id (required), updates (required). Returns: result
Create an account-level service principal.
Inputs: body (required). Returns: result
Delete an account-level service principal.
Inputs: sp_id (required). Returns: result
Fetch an account-level service principal.
Inputs: sp_id (required). Returns: result
List account-level service principals.
Inputs: filter, startIndex, count. Returns: result
Create a metastore storage configuration.
Inputs: body (required). Returns: result
Delete a storage configuration.
Inputs: config_id (required). Returns: result
List metastore storage configurations under the account.
Inputs: no named parameters. Returns: result
Create an account-level SCIM user.
Inputs: body (required). Returns: result
Delete an account-level SCIM user.
Inputs: id (required). Returns: result
Fetch an account-level SCIM user.
Inputs: id (required). Returns: result
List account-level SCIM users.
Inputs: filter, startIndex, count. Returns: result
Return the authenticated account principal via SCIM /Me.
Inputs: no named parameters. Returns: result
Create a new workspace under the account.
Inputs: workspace_name (required), body. Returns: result
Delete a workspace.
Inputs: workspace_id (required). Returns: result
Fetch a workspace by ID.
Inputs: workspace_id (required). Returns: result
List workspaces under the account.
Inputs: no named parameters. Returns: result
Patch a workspace.
Inputs: workspace_id (required), updates (required). Returns: result
Create a new SQL alert. Returns id.
Inputs: name (required), query_id (required), condition (required), notify, +8 more. Returns: result
Delete a SQL alert.
Inputs: id (required). Returns: result
Fetch a single SQL alert definition.
Inputs: id (required). Returns: result
List SQL alerts.
Inputs: page_size, page_token. Returns: result
Update an existing SQL alert.
Inputs: id (required), name (required), query_id (required), condition (required), +9 more. Returns: result
Create a new Databricks App.
Inputs: name (required), description, default_source_code_path, resources. Returns: result
Delete a Databricks App.
Inputs: name (required). Returns: result
Create a deployment for an App.
Inputs: name (required), source_code_path (required), mode. Returns: result
Fetch a single Databricks App.
Inputs: name (required). Returns: result
List all Databricks Apps in the workspace.
Inputs: no named parameters. Returns: result
List deployments for a Databricks App.
Inputs: name (required). Returns: result
Patch an existing Databricks App.
Inputs: name (required), updates (required). Returns: result
Return the resolved Databricks MCP configuration (host, transport, auth method).
Does NOT include any secrets. Useful for verifying configuration.
Inputs: no named parameters. Returns: result
Create a cluster policy. Returns policy_id.
Inputs: name (required), definition (required), description, max_clusters_per_user, +2 more. Returns: result
Delete a cluster policy.
Inputs: policy_id (required). Returns: result
Edit a cluster policy. Returns policy_id.
Inputs: policy_id (required), name, definition, description, +2 more. Returns: result
Fetch a single cluster policy.
Inputs: policy_id (required). Returns: result
List cluster policies.
Inputs: sort_by, order, page_token, page_size, +1 more. Returns: result
Create a Databricks cluster. Returns cluster_id on success.
Use :func:`clusters_spark_versions` and :func:`clusters_list_node_types`
to discover valid values for ``spark_version`` and ``node_type_id``.
Inputs: cluster_name (required), spark_version (required), node_type_id, instance_pool_id, +15 more. Returns: result
Permanently delete a cluster (terminates first).
Inputs: cluster_id (required). Returns: result
Edit a Databricks cluster's configuration. Returns cluster_id.
Inputs: cluster_id (required), cluster_name, num_workers, spark_version, +5 more. Returns: result
List cluster events (ordered from oldest by default).
Inputs: cluster_id (required), limit, offset, order, +1 more. Returns: result
Fetch a cluster's current state and configuration.
Inputs: cluster_id (required). Returns: result
List all clusters visible to the authenticated principal (paged).
Inputs: filter_by, limit, offset. Returns: result
List all available node types for cluster creation.
Inputs: no named parameters. Returns: result
Resize a cluster (must be RUNNING).
Inputs: cluster_id (required), num_workers. Returns: result
Restart a running cluster.
Inputs: cluster_id (required), restart_user. Returns: result
List all available Spark runtime versions.
Inputs: no named parameters. Returns: result
Start a terminated cluster.
Inputs: cluster_id (required). Returns: result
Terminate a cluster (recoverable for 30 days; use :func:clusters_delete to remove permanently).
Inputs: cluster_id (required). Returns: result
Delete a file or directory on DBFS.
Inputs: path (required), recursive. Returns: result
Read bytes from a DBFS file (returned base64-encoded).
Inputs: path (required), offset, length. Returns: result
Get the FileInfo for a DBFS path.
Inputs: path (required). Returns: result
List the contents of a DBFS directory.
Inputs: path (required). Returns: result
Create the given directory (including parents).
Inputs: path (required). Returns: result
Move (rename) a file or directory on DBFS.
Inputs: source_path (required), destination_path (required). Returns: result
Upload contents to a DBFS path.
Inputs: path (required), contents (required), overwrite, encoding. Returns: result
Create an external model (POST /api/2.0/external-models).
Inputs: name (required), provider (required), task (required), config (required). Returns: result
Delete an external model (DELETE /api/2.0/external-models/{name}).
Inputs: name (required). Returns: result
Get an external model (GET /api/2.0/external-models/{name}).
Inputs: name (required). Returns: result
List external models (GET /api/2.0/external-models).
Inputs: no named parameters. Returns: result
Update an external model (PATCH /api/2.0/external-models/{name}).
Inputs: name (required), body (required). Returns: result
Create a Genie AI/BI space.
Inputs: title (required), description, data_source_id, warehouse_id, +1 more. Returns: result
Delete a Genie space.
Inputs: space_id (required). Returns: result
Fetch a single message from a Genie conversation.
Inputs: space_id (required), conversation_id (required), message_id (required). Returns: result
Fetch a Genie space by ID.
Inputs: space_id (required). Returns: result
List conversations in a Genie space.
Inputs: space_id (required). Returns: result
List messages in a Genie conversation.
Inputs: space_id (required), conversation_id (required). Returns: result
List Genie AI/BI spaces the caller can access.
Inputs: no named parameters. Returns: result
Send a follow-up message in an existing Genie conversation.
Inputs: space_id (required), conversation_id (required), content (required), query_parameters. Returns: result
Start a new conversation in a Genie space.
Inputs: space_id (required), question (required). Returns: result
Patch a Genie space.
Inputs: space_id (required), updates (required). Returns: result
Create a SCIM group.
Inputs: displayName (required), members. Returns: result
Delete a SCIM group from the workspace.
Inputs: id (required). Returns: result
Fetch a single SCIM group.
Inputs: id (required). Returns: result
List SCIM groups in the workspace.
Inputs: filter, startIndex, count. Returns: result
Apply partial modifications to a SCIM group.
Inputs: id (required), Operations (required). Returns: result
Replace a SCIM group via PUT.
Inputs: id (required), body (required). Returns: result
Create a new IP access list.
Inputs: label (required), list_type (required), ip_addresses (required), enabled, +1 more. Returns: result
Delete an IP access list.
Inputs: ip_access_list_id (required). Returns: result
Fetch a single IP access list.
Inputs: ip_access_list_id (required). Returns: result
List all IP access lists for the workspace.
Inputs: no named parameters. Returns: result
Update an existing IP access list.
Inputs: ip_access_list_id (required), label, list_type, ip_addresses, +1 more. Returns: result
Create a Databricks job. Returns job_id.
Inputs: name (required), tasks (required), tags, max_concurrent_runs, +13 more. Returns: result
Delete a job (trash, recoverable for 30 days).
Inputs: job_id (required). Returns: result
Fetch a single job's definition.
Inputs: job_id (required). Returns: result
List jobs visible to the authenticated principal.
Inputs: limit, page_token, name, expand_tasks. Returns: result
Fully replace a job's settings (overwrites all fields not in new_settings).
Inputs: job_id (required), new_settings (required), fields_to_remove. Returns: result
Trigger a job run and return the run_id. By default, parameters not listed
are taken from the job's existing parameter defaults.
Inputs: job_id (required), notebook_params, python_params, jar_params, +8 more. Returns: result
Cancel a run (no effect if already completed).
Inputs: run_id (required), timeout_seconds. Returns: result
Cancel all currently active runs of a job.
Inputs: job_id (required). Returns: result
Export and retrieve a run (returns notebook content and dashboard definitions).
Inputs: run_id (required), views_to_export. Returns: result
Get details of a single run, including task list and lifecycle.
Inputs: run_id (required), include_history. Returns: result
Get the output (notebook output, logs, error) of a run. Useful after the run completes.
Inputs: run_id (required). Returns: result
List historical job runs.
Inputs: job_id, active_only, completed_only, limit, +4 more. Returns: result
Re-run a failed run (optionally only specific tasks).
Inputs: run_id (required), rerun_tasks, rerun_all_failed_tasks, latest_repair_id, +1 more. Returns: result
Update a job's settings.
Inputs: job_id (required), new_settings (required), fields_to_remove. Returns: result
Create a Lakeview dashboard.
Inputs: display_name (required), warehouse_id (required), parent, serialized_dashboard. Returns: result
Trash a Lakeview dashboard.
Inputs: dashboard_id (required). Returns: result
Get a Lakeview dashboard's metadata and content.
Inputs: dashboard_id (required). Returns: result
List Lakeview dashboards in the workspace.
Inputs: page_size, page_token, show_trashed. Returns: result
Publish the latest draft of a Lakeview dashboard.
Inputs: dashboard_id (required), embed_credentials. Returns: result
Permanently delete a trashed Lakeview dashboard.
Inputs: dashboard_id (required). Returns: result
Restore a trashed Lakeview dashboard.
Inputs: dashboard_id (required). Returns: result
List trashed Lakeview dashboards.
Inputs: page_size, page_token. Returns: result
Unpublish a Lakeview dashboard.
Inputs: dashboard_id (required). Returns: result
Update a Lakeview dashboard (creates a draft).
Inputs: dashboard_id (required), display_name, warehouse_id, serialized_dashboard. Returns: result
Create a Lakeview dashboard schedule.
Inputs: dashboard_id (required), cron (required), timezone_id (required), paused. Returns: result
Delete a Lakeview dashboard schedule.
Inputs: dashboard_id (required), schedule_id (required). Returns: result
Get a Lakeview dashboard schedule.
Inputs: dashboard_id (required), schedule_id (required). Returns: result
List Lakeview dashboard schedules.
Inputs: dashboard_id (required), page_size, page_token. Returns: result
Update a Lakeview dashboard schedule.
Inputs: dashboard_id (required), schedule_id (required), cron, timezone_id, +1 more. Returns: result
Get a legacy SQL dashboard.
Inputs: dashboard_id (required). Returns: result
List legacy SQL dashboards (/preview/sql/dashboards).
Inputs: page_size, page_token, run_as_role. Returns: result
Create an experiment (POST /api/2.0/mlflow/experiments/create).
Inputs: name (required), artifact_location, tags. Returns: result
Delete an experiment (POST /api/2.0/mlflow/experiments/delete).
Inputs: experiment_id (required). Returns: result
Get an experiment by ID (GET /api/2.0/mlflow/experiments/get).
Inputs: experiment_id (required). Returns: result
Search experiments (GET /api/2.0/mlflow/experiments/search).
Inputs: max_results, page_token, filter, order_by, +1 more. Returns: result
Update an experiment (POST /api/2.0/mlflow/experiments/update).
Inputs: experiment_id (required), new_name, tags, lifecycle_stage. Returns: result
Create a model version (POST /api/2.0/mlflow/model-versions/create).
Inputs: name (required), source (required), run_id, tags, +2 more. Returns: result
Delete a model version (POST /api/2.0/mlflow/model-versions/delete).
Inputs: name (required), version (required). Returns: result
Get a model version (GET /api/2.0/mlflow/model-versions/get).
Inputs: name (required), version (required). Returns: result
Search model versions (POST /api/2.0/mlflow/model-versions/search).
Inputs: filter, max_results, order_by, page_token. Returns: result
Transition a model version's stage
(POST /api/2.0/mlflow/model-versions/transition-stage).
Inputs: name (required), version (required), stage (required), archive_existing_versions. Returns: result
Update a model version (POST /api/2.0/mlflow/model-versions/update).
Inputs: name (required), version (required), description. Returns: result
Create a registered model (POST /api/2.0/mlflow/registered-models/create).
Inputs: name (required), tags, description. Returns: result
Delete a registered model (POST /api/2.0/mlflow/registered-models/delete).
Inputs: name (required). Returns: result
Get a registered model (GET /api/2.0/mlflow/registered-models/get).
Inputs: name (required). Returns: result
Search registered models (POST /api/2.0/mlflow/registered-models/search).
Inputs: filter, max_results, order_by, page_token. Returns: result
Update a registered model (POST /api/2.0/mlflow/registered-models/update).
Inputs: name (required), description, new_name. Returns: result
Create a run (POST /api/2.0/mlflow/runs/create).
Inputs: experiment_id (required), start_time, run_name, tags, +1 more. Returns: result
Delete a run (POST /api/2.0/mlflow/runs/delete).
Inputs: run_id (required). Returns: result
Delete a tag (POST /api/2.0/mlflow/runs/delete-tag).
Inputs: run_id (required), key (required). Returns: result
Get a run (GET /api/2.0/mlflow/runs/get).
Inputs: run_id (required), run_uuid. Returns: result
List run artifacts (GET /api/2.0/mlflow/runs/list-artifacts).
Inputs: run_id (required), path. Returns: result
Log many metrics/params/tags in one call
(POST /api/2.0/mlflow/runs/log-batch).
Inputs: run_id (required), metrics, params, tags. Returns: result
Log a single metric (POST /api/2.0/mlflow/runs/log-metric).
Inputs: run_id (required), key (required), value (required), timestamp, +1 more. Returns: result
Log a single param (POST /api/2.0/mlflow/runs/log-param).
Inputs: run_id (required), key (required), value (required). Returns: result
Search runs (POST /api/2.0/mlflow/runs/search).
Inputs: experiment_ids, filter, max_results, order_by, +2 more. Returns: result
Set a tag (POST /api/2.0/mlflow/runs/set-tag).
Inputs: run_id (required), key (required), value (required). Returns: result
Update a run (POST /api/2.0/mlflow/runs/update).
Inputs: run_id (required), status, end_time, run_name. Returns: result
Delete ACL entries from an object (or all of them when none provided).
Inputs: object_type (required), object_id (required), access_control_list. Returns: result
Fetch the ACL for the given object.
Inputs: object_type (required), object_id (required). Returns: result
Return the valid permission levels for the given object type.
Inputs: object_type (required). Returns: result
Resolve ACLs for a list of object IDs (POST /{object_type}/objectIds).
Inputs: object_type (required), object_ids (required). Returns: result
Replace the ACL for an object (POST replaces the entire ACL).
Inputs: object_type (required), object_id (required), access_control_list (required). Returns: result
Patch the ACL — entries here are merged into the existing ACL.
Inputs: object_type (required), object_id (required), access_control_list (required). Returns: result
Create a pipeline (POST /api/2.0/pipelines).
Inputs: name, storage, configuration, cluster, +15 more. Returns: result
Delete a pipeline (DELETE /api/2.0/pipelines/{pipeline_id}).
Inputs: pipeline_id (required). Returns: result
Edit a pipeline (PUT /api/2.0/pipelines/{pipeline_id}).
Inputs: pipeline_id (required), name, storage, configuration, +15 more. Returns: result
Get a pipeline (GET /api/2.0/pipelines/{pipeline_id}).
Inputs: pipeline_id (required). Returns: result
Get a specific pipeline update
(GET /api/2.0/pipelines/{pipeline_id}/updates/{update_id}).
Inputs: pipeline_id (required), update_id (required). Returns: result
List pipelines (GET /api/2.0/pipelines).
Inputs: max_results, page_token, filter. Returns: result
List pipeline events
(GET /api/2.0/pipelines/{pipeline_id}/events).
Inputs: pipeline_id (required), order, page_token, max_results, +2 more. Returns: result
List updates for a pipeline
(GET /api/2.0/pipelines/{pipeline_id}/updates).
Inputs: pipeline_id (required), page_token, max_results. Returns: result
Start a pipeline update
(POST /api/2.0/pipelines/{pipeline_id}/updates).
Inputs: pipeline_id (required), full_refresh, refresh_selection, rewind_spec, +3 more. Returns: result
Stop a pipeline (POST /api/2.0/pipelines/{pipeline_id}/stop).
Inputs: pipeline_id (required). Returns: result
Create a new instance pool.
Inputs: instance_pool_name (required), node_type_id, min_idle_instances, max_capacity, +11 more. Returns: result
Delete an instance pool.
Inputs: instance_pool_id (required). Returns: result
Edit an existing instance pool. Returns instance_pool_id.
Inputs: instance_pool_id (required), instance_pool_name, node_type_id, min_idle_instances, +7 more. Returns: result
Fetch a single instance pool.
Inputs: instance_pool_id (required). Returns: result
List all instance pools in the workspace.
Inputs: no named parameters. Returns: result
Create a new saved SQL query. Returns id.
Inputs: name (required), query_text (required), data_source_id (required), description, +7 more. Returns: result
Delete a saved SQL query (trash).
Inputs: id (required). Returns: result
Fetch a single saved SQL query definition.
Inputs: id (required). Returns: result
List saved SQL queries.
Inputs: q, tags, page_size, page_token, +2 more. Returns: result
Update a saved SQL query.
Inputs: id (required), name (required), query_text (required), data_source_id (required), +8 more. Returns: result
List historical SQL query executions with optional filters.
Builds the ``filter_by`` payload for the Databricks query history API
from the supplied keyword arguments. ``include_metrics`` and pagination
parameters are passed as top-level query string parameters.
Inputs: start_time_ms, end_time_ms, user_ids, statements, +8 more. Returns: result
Create (and clone) a new repo in the workspace.
Inputs: url (required), provider (required), path, branch, +2 more. Returns: result
Delete a workspace repo.
Inputs: repo_id (required). Returns: result
Fetch a single repo.
Inputs: repo_id (required). Returns: result
List workspace repos (cursor-paginated).
Inputs: path_prefix, next_page_token, page_size. Returns: result
Update a repo — switch branch/tag or change sparse checkout.
Inputs: repo_id (required), branch, tag, repo_path, +1 more. Returns: result
Create a new secret scope.
Inputs: scope (required), initial_manage_principal, backend_type, keyvault_metadata. Returns: result
Remove an ACL from a secret scope.
Inputs: scope (required), principal (required). Returns: result
Delete a secret scope (all secrets within must already be deleted).
Inputs: scope (required). Returns: result
Delete the secret stored at key within scope.
Inputs: scope (required), key (required). Returns: result
Fetch the ACL applied to principal on a secret scope.
Inputs: scope (required), principal (required). Returns: result
Retrieve a secret value. Returns the decoded secret string (or base64 for bytes).
Inputs: scope (required), key (required). Returns: result
List the secret keys within a scope (values are NOT returned).
Inputs: scope (required). Returns: result
List ACLs for the given secret scope.
Inputs: scope (required). Returns: result
List all secret scopes visible to the authenticated principal.
Inputs: no named parameters. Returns: result
Put a secret into the given scope. Returns {} on success.
Inputs: scope (required), key (required), string_value, bytes_value. Returns: result
Create or overwrite an ACL on a secret scope.
Inputs: scope (required), principal (required), permission (required). Returns: result
Create a SCIM service principal.
Inputs: displayName (required), applicationId, entitlements, groups, +2 more. Returns: result
Delete a SCIM service principal from the workspace.
Inputs: id (required). Returns: result
Fetch a single SCIM service principal.
Inputs: id (required). Returns: result
List SCIM service principals in the workspace.
Inputs: filter, startIndex, count. Returns: result
Apply partial modifications to a SCIM service principal.
Inputs: id (required), Operations (required). Returns: result
Replace a SCIM service principal via PUT.
Inputs: id (required), body (required). Returns: result
Create a serving endpoint (POST /api/2.0/serving-endpoints).
Inputs: name (required), config (required), route_optimized. Returns: result
Delete a serving endpoint (DELETE /api/2.0/serving-endpoints/{name}).
Inputs: name (required). Returns: result
Get a serving endpoint (GET /api/2.0/serving-endpoints/{name}).
Inputs: name (required). Returns: result
List model serving endpoints (GET /api/2.0/serving-endpoints).
Inputs: page_size, page_token. Returns: result
Fetch the OpenAPI schema for a serving endpoint
(GET /api/2.0/serving-endpoints/{name}/openapi).
Inputs: name (required). Returns: result
Query/invoke a serving endpoint
(POST /api/2.0/serving-endpoints/{name}/invocations).
Inputs: name (required), body (required). Returns: result
Update a serving endpoint config
(PATCH /api/2.0/serving-endpoints/{name}).
Inputs: name (required), config (required). Returns: result
Create a Delta Sharing provider
(POST /api/2.1/unity-catalog/providers).
Inputs: name (required), recipient_profile_str, comment, authentication_type, +1 more. Returns: result
Delete a Delta Sharing provider
(DELETE /api/2.1/unity-catalog/providers/{name}).
Inputs: name (required). Returns: result
Get a Delta Sharing provider
(GET /api/2.1/unity-catalog/providers/{name}).
Inputs: name (required). Returns: result
List Delta Sharing providers
(GET /api/2.1/unity-catalog/providers).
Inputs: no named parameters. Returns: result
Update a Delta Sharing provider
(PATCH /api/2.1/unity-catalog/providers/{name}).
Inputs: name (required), body (required). Returns: result
Create a Delta Sharing recipient
(POST /api/2.1/unity-catalog/recipients).
Inputs: name (required), comment, authentication_type, sharing_code, +2 more. Returns: result
Delete a Delta Sharing recipient
(DELETE /api/2.1/unity-catalog/recipients/{name}).
Inputs: name (required). Returns: result
Get a Delta Sharing recipient
(GET /api/2.1/unity-catalog/recipients/{name}).
Inputs: name (required). Returns: result
List Delta Sharing recipients
(GET /api/2.1/unity-catalog/recipients).
Inputs: no named parameters. Returns: result
Rotate the token for a Delta Sharing recipient
(POST /api/2.1/unity-catalog/recipients/{name}/rotate-token).
Inputs: name (required). Returns: result
Update a Delta Sharing recipient
(PATCH /api/2.1/unity-catalog/recipients/{name}).
Inputs: name (required), body (required). Returns: result
Cancel a running SQL statement.
Inputs: statement_id (required). Returns: result
Execute a SQL statement against a warehouse. Returns the statement statement_id and result.
Poll :func:`sql_statements_get` (or fetch a chunk via :func:`sql_statements_get_result_chunk`)
to retrieve results once the statement finishes.
Inputs: statement (required), warehouse_id (required), catalog, schema, +8 more. Returns: result
Get a statement's status and (when ready) its inline result manifest.
Inputs: statement_id (required). Returns: result
Fetch a single chunk of an EXTERNAL_LINKS result.
Inputs: statement_id (required), chunk_index (required). Returns: result
Create a new PAT token. Returns token_value (only returned once).
Inputs: lifetime_seconds, comment. Returns: result
Return metadata for the current PAT token (the one used to authenticate).
Inputs: no named parameters. Returns: result
List all PAT tokens owned by the authenticated principal.
Inputs: no named parameters. Returns: result
Revoke (delete) a PAT token by ID.
Inputs: token_id (required). Returns: result
Create a catalog (POST /api/2.1/unity-catalog/catalogs).
Inputs: name (required), comment, properties, storage_root, +9 more. Returns: result
Delete a catalog (DELETE /api/2.1/unity-catalog/catalogs/{name}?force=...).
Inputs: name (required), force. Returns: result
Get a single catalog (GET /api/2.1/unity-catalog/catalogs/{name}).
Inputs: name (required). Returns: result
List catalogs (GET /api/2.1/unity-catalog/catalogs).
Inputs: max_results, page_token, include_browse, include_managed_properties, +1 more. Returns: result
Update a catalog (PATCH /api/2.1/unity-catalog/catalogs/{name}).
Inputs: name (required), comment, properties, owner, +2 more. Returns: result
Create a function (POST /api/2.1/unity-catalog/functions).
Inputs: name (required), catalog_name (required), schema_name (required), input_params, +9 more. Returns: result
Delete a function (DELETE /api/2.1/unity-catalog/functions/{full_name}).
Inputs: full_name (required). Returns: result
Get a single function (GET /api/2.1/unity-catalog/functions/{full_name}).
Inputs: full_name (required). Returns: result
List functions (GET /api/2.1/unity-catalog/functions).
Inputs: catalog_name, schema_name, max_results, page_token. Returns: result
Update a function (PATCH /api/2.1/unity-catalog/functions/{full_name}).
Inputs: full_name (required), body (required). Returns: result
Get effective permissions on a securable
(GET /api/2.1/unity-catalog/permissions/{securable_type}/{full_name}).
Inputs: securable_type (required), full_name (required). Returns: result
Apply permission changes to a securable
(PATCH /api/2.1/unity-catalog/permissions/{securable_type}/{full_name}).
Inputs: securable_type (required), full_name (required), changes (required). Returns: result
Create a registered model (POST /api/2.1/unity-catalog/models).
Inputs: name (required), catalog_name (required), schema_name (required), comment, +6 more. Returns: result
Delete a registered model
(DELETE /api/2.1/unity-catalog/models/{full_name}).
Inputs: full_name (required). Returns: result
Get a registered model (GET /api/2.1/unity-catalog/models/{full_name}).
Inputs: full_name (required). Returns: result
List registered models (GET /api/2.1/unity-catalog/models).
Inputs: catalog_name, schema_name, max_results, page_token. Returns: result
Update a registered model (PATCH /api/2.1/unity-catalog/models/{full_name}).
Inputs: full_name (required), comment, new_name, owner, +3 more. Returns: result
Create a model version
(POST /api/2.1/unity-catalog/models/{full_name}/versions).
Inputs: full_name (required), source (required), run_id, description, +4 more. Returns: result
Delete a model version
(DELETE /api/2.1/unity-catalog/models/{full_name}/versions/{version}).
Inputs: full_name (required), version (required). Returns: result
Delete a model alias
(DELETE /api/2.1/unity-catalog/models/{full_name}/aliases/{alias}).
Inputs: full_name (required), alias (required). Returns: result
Get a model version by number or alias
(GET /api/2.1/unity-catalog/models/{full_name}/versions/{version_or_alias}).
Inputs: full_name (required), version (required). Returns: result
Set (or update) a model alias to point at a version
(PATCH /api/2.1/unity-catalog/models/{full_name}/aliases/{alias}).
Inputs: full_name (required), alias (required), version (required). Returns: result
List model versions
(GET /api/2.1/unity-catalog/models/{full_name}/versions).
Inputs: full_name (required), max_results, page_token, include_browse, +1 more. Returns: result
Create a schema (POST /api/2.1/unity-catalog/schemas).
Inputs: name (required), catalog_name (required), comment, properties, +4 more. Returns: result
Delete a schema (DELETE /api/2.1/unity-catalog/schemas/{full_name}).
Inputs: full_name (required). Returns: result
Get a single schema (GET /api/2.1/unity-catalog/schemas/{full_name}).
Inputs: full_name (required). Returns: result
List schemas (GET /api/2.1/unity-catalog/schemas).
Inputs: catalog_name, max_results, page_token, include_browse, +2 more. Returns: result
Update a schema (PATCH /api/2.1/unity-catalog/schemas/{full_name}).
Inputs: full_name (required), comment, properties, owner, +2 more. Returns: result
Create a table (POST /api/2.1/unity-catalog/tables).
Inputs: name (required), catalog_name (required), schema_name (required), table_type (required), +14 more. Returns: result
Delete a table (DELETE /api/2.1/unity-catalog/tables/{full_name}).
Inputs: full_name (required). Returns: result
Get a single table (GET /api/2.1/unity-catalog/tables/{full_name}).
Inputs: full_name (required), include_delta_metadata, include_browse, include_managed_properties, +2 more. Returns: result
List tables (GET /api/2.1/unity-catalog/tables).
Inputs: catalog_name, schema_name, table_names, max_results, +8 more. Returns: result
Update a table (PATCH /api/2.1/unity-catalog/tables/{full_name}).
Inputs: full_name (required), body (required). Returns: result
Create a volume (POST /api/2.1/unity-catalog/volumes).
Inputs: name (required), catalog_name (required), schema_name (required), volume_type (required), +5 more. Returns: result
Delete a volume (DELETE /api/2.1/unity-catalog/volumes/{full_name}).
Inputs: full_name (required). Returns: result
Get a single volume (GET /api/2.1/unity-catalog/volumes/{full_name}).
Inputs: full_name (required). Returns: result
List volumes (GET /api/2.1/unity-catalog/volumes).
Inputs: catalog_name, schema_name, max_results, page_token. Returns: result
Update a volume (PATCH /api/2.1/unity-catalog/volumes/{full_name}).
Inputs: full_name (required), body (required). Returns: result
Create a new SCIM user.
Inputs: userName (required), name, emails, active, +2 more. Returns: result
Delete a SCIM user from the workspace.
Inputs: id (required). Returns: result
Fetch a single SCIM user.
Inputs: id (required). Returns: result
List SCIM users in the workspace.
Inputs: filter, startIndex, count. Returns: result
Apply partial modifications to a SCIM user.
Inputs: id (required), Operations (required). Returns: result
Replace a user via PUT (must include all required fields).
Inputs: id (required), body (required). Returns: result
Create a Vector Search endpoint
(POST /api/2.0/vector-search/endpoints).
Inputs: name (required), endpoint_type (required). Returns: result
Delete a Vector Search endpoint
(DELETE /api/2.0/vector-search/endpoints/{endpoint_id}).
Inputs: endpoint_id (required). Returns: result
Get a Vector Search endpoint
(GET /api/2.0/vector-search/endpoints/{endpoint_id}).
Inputs: endpoint_id (required). Returns: result
List Vector Search endpoints (GET /api/2.0/vector-search/endpoints).
Inputs: no named parameters. Returns: result
Create a Vector Search index
(POST /api/2.0/vector-search/indexes).
Inputs: name (required), endpoint_name (required), primary_key (required), index_type (required), +5 more. Returns: result
Delete a Vector Search index
(DELETE /api/2.0/vector-search/indexes/{index_name}).
Inputs: index_name (required). Returns: result
Get a Vector Search index
(GET /api/2.0/vector-search/indexes/{index_name}).
Inputs: index_name (required). Returns: result
List Vector Search indexes
(GET /api/2.0/vector-search/indexes).
Inputs: endpoint_name. Returns: result
Trigger a sync for a Delta-sync index
(POST /api/2.0/vector-search/indexes/{index_name}/sync).
Inputs: index_name (required). Returns: result
Query a Vector Search index
(POST /api/2.0/vector-search/query-index).
Inputs: index_name (required), query_text, query_vector, columns, +6 more. Returns: result
Create a new SQL warehouse. Returns the created warehouse's id.
Inputs: name (required), cluster_size (required), max_num_clusters, min_num_clusters, +16 more. Returns: result
Permanently delete a SQL warehouse.
Inputs: id (required). Returns: result
Edit an existing SQL warehouse's configuration. Returns the warehouse id.
Inputs: id (required), name, cluster_size, max_num_clusters, +16 more. Returns: result
Fetch a single SQL warehouse's configuration and state.
Inputs: id (required). Returns: result
List all SQL warehouses visible to the authenticated principal.
Returns an empty list when there are no warehouses.
Inputs: no named parameters. Returns: result
Start a SQL warehouse.
Inputs: id (required). Returns: result
Stop a running SQL warehouse.
Inputs: id (required). Returns: result
Return the authenticated principal (Databricks /api/2.0/preview/scim/v2/Me).
Inputs: no named parameters. Returns: result
Return configuration for a specific set of keys.
Inputs: keys (required). Returns: result
Return the full workspace configuration (all keys/values).
Inputs: no named parameters. Returns: result
Set (or unset) workspace configuration values via PATCH /workspace-conf.
Inputs: keys (required), remove. Returns: result
Delete a workspace object (notebook, file, directory).
Inputs: path (required), recursive. Returns: result
Export a workspace object (notebook or file).
For notebooks, ``format_`` may be ``SOURCE``, ``HTML``, ``JUPYTER``, or ``DBC``.
For files, use ``RAW`` (returns base64) or omit.
Returns the decoded text where possible.
Inputs: path (required), format_. Returns: result
Get the status (object_type, path, language) of a workspace object.
Inputs: path (required). Returns: result
Import a workspace object (notebook or file).
Inputs: path (required), content (required), format_, language, +1 more. Returns: result
List the contents of a workspace directory.
Inputs: path, limit, offset. Returns: result
List files under /Workspace/files/<directory_path>.
Inputs: directory_path (required). Returns: result
Create a workspace directory (and any necessary parents).
Inputs: path (required). Returns: result
Read the text content of a small file under /Workspace/files.
Inputs: file_path (required). Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 303 tools
Resource-prefixed names make most tools easy to tell apart, with each domain (clusters, jobs, UC, MLflow, etc.) following clear CRUD patterns. However, overlaps like uc_model vs mlflow_registered_models, workspace_list vs dbfs_list, and whoami vs account_whoami create some ambiguity in a 303-tool surface.
The dominant pattern is resource_subresource_verb (e.g., clusters_list, uc_table_create, sharing_providers_get), which is highly predictable. Minor deviations such as whoami, auth_config, jobs_run_now, vs_query_index, and clusters_list_node_types prevent a perfect score.
303 tools is an extreme mismatch for an agent-facing MCP surface; even though Databricks is a broad platform, no agent can effectively select from this many tools. This far exceeds the 'too many' threshold and significantly harms usability.
The tool surface is remarkably comprehensive, covering CRUD/lifecycle operations across clusters, jobs, warehouses, SQL, UC, MLflow, serving, sharing, secrets, permissions, SCIM, repos, DBFS, pools, apps, and account management. Minor gaps exist (e.g., account-level SCIM lacks update/patch, account storage configs lack get/update, legacy dashboards are read-only) but they are workarounds.
Governed data discovery, exact queries, decisions, simulations, and runtime utilities over MCP.
Public tools to understand Dynamik, discover datasets, and connect account-scoped capabilities.
Your org's AI agents, tasks, runs, search, and brain files as MCP tools and resources.
