Provides tools for workspace Git integration in Microsoft Fabric, enabling version control management of Fabric items.
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., "@Microsoft Fabric MCP Server list workspaces"
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 production-grade Model Context Protocol server for Microsoft Fabric. Exposes Fabric operations as MCP tools over stdio (local) and Streamable HTTP (hosted/multi-user).
TypeScript (ESM, Node 20+), REST-only (Fabric has no typed SDK)
Transparent Long-Running Operation (LRO) handling — 202 + Location is polled to completion automatically
Three auth modes: Azure CLI (az login), Service Principal (deploy path — Fabric has no PAT), OBO + Entra JWT (hosted)
Per-tool RBAC, destructive-op guard, secret redaction, structured logging, retry + per-host circuit breaker
Wraps the Fabric REST API, the Power BI API, and OneLake (per-resource token scopes)
Status: All phases complete — 105 tools across 17 domains, plus 4 resources and 4 prompts:
core (11) · items (7) · lakehouses (7) · warehouses (4) · sqlendpoints (3) · shortcuts (4)
notebooks (6) · pipelines (6) · jobs (6) · spark (6)
semanticmodels (5) · reports (3) · dataflows (5)
realtime (12: eventhouses, KQL databases, eventstreams) · onelake (5: files/dirs via DFS)
git (7: workspace Git integration) · platform (8: deployment pipelines, connections, gateways)
Resources:
fabric://workspaces,fabric://workspace/{w}/items,…/lakehouses,…/item/{id}Prompts:
medallion-architecture,document-workspace,build-etl,diagnose-refreshHardening: per-host circuit breaker, retry + Retry-After, transparent LRO polling, secret redaction, per-tool RBAC + destructive guard, audit logging, ESLint, Vitest suite (unit + live MCP-protocol test), Azure Pipelines CI.
Node.js ≥ 20
A Fabric-enabled tenant and an identity with a Fabric/Power BI license (or a service principal enabled for Fabric APIs and added to the workspaces)
One of: Azure CLI (az login), a service principal, or (hosted) an Entra app for OBO
Related MCP server: powerbi-mcp
cd "E:\MCP Servers\fabric-mcp-server"
npm install
cp .env.example .env # PowerShell: Copy-Item .env.example .envMinimum for local use with the Azure CLI:
AUTH_MODE=azcli
FABRIC_DEFAULT_WORKSPACE= # optional default workspace ID
| Needs | Notes |
|
| Easiest for dev. Token minted for the Fabric scope. |
|
| Deployment path. The SP must be enabled in Fabric tenant settings and added to the workspaces. |
| same three | HTTP transport only. Validates the caller's Entra JWT, then exchanges it per resource (preserves user identity). |
npm run build && npm start # stdio (default)
node dist/index.js --transport http # Streamable HTTP on PORT (default 3100){
"mcpServers": {
"fabric": {
"command": "npx",
"args": ["-y", "fabric-mcp-server"],
"env": { "AUTH_MODE": "azcli", "FABRIC_DEFAULT_WORKSPACE": "<workspace-id>" }
}
}
}For deployment use AUTH_MODE=sp with the three AZURE_* env vars.
Script | Purpose |
| Compile TypeScript → |
| Run compiled server (stdio / HTTP) |
| Watch-mode dev (tsx) |
| Quality gates |
transport (stdio | http)
→ server (McpServer + registry)
→ dispatch (RBAC · destructive guard · audit · error envelope)
→ domain tools (Zod schemas)
→ domain services
→ FabricClient (REST + retry + circuit breaker + LRO polling)
→ credential (AzCli | ServicePrincipal | OBO) → per-resource tokenEach Fabric area is a self-contained module under src/domains/<area>/ exporting a DomainModule. To add an area, create the module and append it to domains in src/server.ts.
Key Fabric specifics: every write is an LRO (handled in src/client/fabricClient.ts); item definitions (notebooks, pipelines) are base64 file parts (helpers in src/domains/_shared.ts); tokens are acquired per resource (Fabric / Power BI / OneLake).
Full walkthrough: see docs/END_TO_END_FLOW.md for how a request travels through every layer (startup, a tool call step by step, the LRO + item-definition mechanisms, auth/scope flows, resources/prompts, and a worked medallion-ETL example).
Add a member (user/group/service principal) to a workspace with a role (Admin/Member/Contributor/Viewer).
Inputs: principalId (required), principalType (required), role (required), workspace.
Assign a workspace to a capacity.
Inputs: capacityId (required), workspace.
Cancel a running job instance.
Inputs: itemId (required), jobInstanceId (required), workspace.
Create a Dataflow Gen2.
Inputs: displayName (required), description, workspace.
Create an eventhouse.
Inputs: displayName (required), description, workspace.
Create an eventstream.
Inputs: displayName (required), description, workspace.
Create any Fabric item by type, optionally with a definition (file parts). Use the type-specific tools for richer ergonomics.
Inputs: displayName (required), type (required), description, definitionParts, +1 more.
Create a schedule for an item's job. configuration is the Fabric schedule object, e.g. {"type":"Daily","startDateTime":"2026-01-01T00:00:00","times":["09:00"],"localTimeZoneId":"UTC"}.
Inputs: itemId (required), jobType (required), configuration (required), enabled, +1 more.
Create a KQL database under an eventhouse.
Inputs: displayName (required), eventhouseId (required), description, workspace.
Create a lakehouse, optionally schema-enabled.
Inputs: displayName (required), description, enableSchemas, workspace.
Create a notebook. Optionally populate cells at creation and attach a default lakehouse.
Inputs: displayName (required), description, cells, defaultLakehouse, +1 more.
Create a data pipeline, optionally with activities + parameters.
Inputs: displayName (required), description, activities, parameters, +1 more.
Create a OneLake shortcut. target is the source definition, e.g. {"oneLake":{"workspaceId":"...","itemId":"...","path":"Tables/x"}} or {"adlsGen2":{...}} / {"amazonS3":{...}}.
Inputs: itemId (required), path (required), name (required), target (required), +1 more.
Create a Spark job definition.
Inputs: displayName (required), description, workspace.
Create a data warehouse.
Inputs: displayName (required), description, workspace.
Create a new workspace, optionally bound to a capacity.
Inputs: displayName (required), description, capacityId.
Delete a connection. Destructive.
Inputs: id (required).
Delete a Dataflow Gen2. Destructive.
Inputs: dataflowId (required), workspace.
Delete an eventhouse. Destructive.
Inputs: id (required), workspace.
Delete an eventstream. Destructive.
Inputs: id (required), workspace.
Delete an item by ID. Destructive.
Inputs: itemId (required), workspace.
Delete a KQL database. Destructive.
Inputs: id (required), workspace.
Delete a lakehouse. Destructive.
Inputs: lakehouseId (required), workspace.
Delete a report. Destructive.
Inputs: reportId (required), workspace.
Delete a semantic model. Destructive.
Inputs: modelId (required), workspace.
Delete a OneLake shortcut by path + name. Destructive.
Inputs: itemId (required), shortcutPath (required), name (required), workspace.
Delete a warehouse. Destructive.
Inputs: warehouseId (required), workspace.
Delete a workspace. Destructive.
Inputs: workspace.
Remove a role assignment from a workspace. Destructive.
Inputs: assignmentId (required), workspace.
Deploy content from one deployment-pipeline stage to the next (long-running). Destructive — promotes content.
Inputs: id (required), sourceStageId (required), targetStageId (required), note, +1 more.
Get a connection by ID.
Inputs: id (required).
Get a Dataflow Gen2 by ID.
Inputs: dataflowId (required), workspace.
Get a deployment pipeline by ID.
Inputs: id (required).
Get the stages of a deployment pipeline.
Inputs: id (required).
Get an eventhouse by ID.
Inputs: id (required), workspace.
Get an eventstream by ID.
Inputs: id (required), workspace.
Get a single item by ID.
Inputs: itemId (required), workspace.
Fetch an item's full definition (long-running). Returns the raw definition and decoded file contents.
Inputs: itemId (required), format, workspace.
Get the status/details of a job instance.
Inputs: itemId (required), jobInstanceId (required), workspace.
Get a KQL database by ID.
Inputs: id (required), workspace.
Get a lakehouse (includes its SQL endpoint + OneLake paths).
Inputs: lakehouseId (required), workspace.
Get a notebook by ID.
Inputs: notebookId (required), workspace.
Fetch a notebook's definition and decoded cell source (long-running).
Inputs: notebookId (required), workspace.
Get a data pipeline by ID.
Inputs: pipelineId (required), workspace.
Fetch a pipeline's definition (activities/parameters) decoded (long-running).
Inputs: pipelineId (required), workspace.
Get a report by ID.
Inputs: reportId (required), workspace.
Get a semantic model by ID.
Inputs: modelId (required), workspace.
Get the refresh history of a semantic model.
Inputs: modelId (required), workspace.
Get a single OneLake shortcut by path + name.
Inputs: itemId (required), shortcutPath (required), name (required), workspace.
Get a Spark job definition by ID.
Inputs: id (required), workspace.
Get a SQL analytics endpoint (connection string, etc.).
Inputs: sqlEndpointId (required), workspace.
Get a warehouse (includes its SQL connection string).
Inputs: warehouseId (required), workspace.
Get a workspace by ID.
Inputs: workspace.
Commit workspace changes to Git (mode All or Selective).
Inputs: mode, comment, items, workspaceHead, +1 more.
Connect a workspace to a Git repository. gitProviderDetails = { gitProviderType: "AzureDevOps"|"GitHub", organizationName, projectName, repositoryName, branchName, directoryName }.
Inputs: gitProviderDetails (required), workspace.
Disconnect a workspace from Git. Destructive.
Inputs: workspace.
Get a workspace's Git connection.
Inputs: workspace.
Get the Git status (changes) of a workspace.
Inputs: workspace.
Initialize the Git connection (after connect) to sync direction.
Inputs: workspace.
Update the workspace from the remote Git commit.
Inputs: remoteCommitHash (required), conflictResolutionPolicy, workspaceHead, workspace.
List the capacities the caller can access.
Inputs: no named parameters.
List connections.
Inputs: no named parameters.
List Dataflow Gen2 items in a workspace.
Inputs: workspace.
List deployment pipelines.
Inputs: no named parameters.
List eventhouses in a workspace.
Inputs: workspace.
List eventstreams in a workspace.
Inputs: workspace.
List gateways.
Inputs: no named parameters.
List items in a workspace, optionally filtered by type (Lakehouse, Notebook, DataPipeline, Warehouse, SemanticModel, Report, …).
Inputs: workspace, type.
List job instances (runs) for an item.
Inputs: itemId (required), workspace.
List schedules configured for an item's job type.
Inputs: itemId (required), jobType (required), workspace.
List KQL databases in a workspace.
Inputs: workspace.
List lakehouses in a workspace.
Inputs: workspace.
List the Delta tables in a lakehouse.
Inputs: lakehouseId (required), workspace.
List Spark Livy sessions in a workspace.
Inputs: workspace.
List notebooks in a workspace.
Inputs: workspace.
List data pipelines in a workspace.
Inputs: workspace.
List reports in a workspace.
Inputs: workspace.
List semantic models (datasets) in a workspace.
Inputs: workspace.
List OneLake shortcuts on an item (e.g. a lakehouse).
Inputs: itemId (required), workspace.
List the custom Spark pools in a workspace.
Inputs: workspace.
List Spark job definitions in a workspace.
Inputs: workspace.
List SQL analytics endpoints in a workspace.
Inputs: workspace.
List warehouses in a workspace.
Inputs: workspace.
List role assignments (members) on a workspace.
Inputs: workspace.
List all Fabric workspaces the caller can access.
Inputs: no named parameters.
Load a file/folder in OneLake into a Delta table (long-running).
Inputs: lakehouseId (required), tableName (required), relativePath (required), pathType (required), +4 more.
Create a directory in OneLake.
Inputs: path (required), workspace.
Delete a file or directory in OneLake. Destructive.
Inputs: path (required), recursive, workspace.
List files/directories under a OneLake directory (e.g. "{lakehouseId}/Files").
Inputs: directory (required), recursive, workspace.
Read a text file from OneLake.
Inputs: path (required), workspace.
Write a (small) text file to OneLake (create → append → flush).
Inputs: path (required), content (required), workspace.
Trigger an asynchronous refresh of a semantic model.
Inputs: modelId (required), workspace.
Sync a SQL analytics endpoint's metadata with its source lakehouse (long-running).
Inputs: sqlEndpointId (required), workspace.
Run (refresh) a Dataflow Gen2 on demand.
Inputs: dataflowId (required), workspace.
Run an on-demand job for any item, by job type. Optionally pass executionData.
Inputs: itemId (required), jobType (required), executionData, workspace.
Run a notebook on demand. Optionally pass executionData (parameters, configuration).
Inputs: notebookId (required), executionData, workspace.
Run a data pipeline on demand. Optionally pass executionData (parameters).
Inputs: pipelineId (required), executionData, workspace.
Run a Spark job definition on demand.
Inputs: id (required), workspace.
Run table maintenance (OPTIMIZE / V-Order / VACUUM) on a lakehouse table (long-running).
Inputs: lakehouseId (required), tableName (required), schemaName, optimize, +3 more.
Unassign a workspace from its capacity.
Inputs: workspace.
Update an item's name/description.
Inputs: itemId (required), displayName, description, workspace.
Replace an item's definition from file parts (long-running).
Inputs: itemId (required), parts (required), workspace.
Replace a notebook's cells (long-running). Optionally attach a default lakehouse.
Inputs: notebookId (required), cells (required), defaultLakehouse, workspace.
Replace a pipeline's activities + parameters (long-running).
Inputs: pipelineId (required), activities (required), parameters, workspace.
Update a workspace name/description.
Inputs: workspace, displayName, description.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 105 tools
The generic item/job tools (fabric_get_item, fabric_delete_item, fabric_run_job, fabric_get_item_definition) overlap directly with type-specific tools (fabric_get_lakehouse, fabric_delete_lakehouse, fabric_run_notebook, fabric_get_notebook_definition), so an agent can easily select the wrong one for the same resource and action. Descriptions help in some cases, but the boundaries between generic and type-specific tools are repeatedly unclear.
Tools consistently use the fabric_ prefix and mostly follow a verb_noun snake_case pattern (list/get/create/update/delete/run). Minor inconsistencies exist, such as domain-prefixed names like fabric_onelake_* and fabric_git_*, add vs create for role assignments, and generic vs type-specific definition/run tools.
105 tools is an extreme count for an MCP tool surface, far beyond even the 25+ threshold. Many tools duplicate generic item operations (list_items vs list_lakehouses, delete_item vs delete_lakehouse, etc.), which inflates the count without adding proportional coherence.
The server covers most major Fabric domains—workspaces, lakehouses, warehouses, notebooks, pipelines, semantic models, shortcuts, Git, deployment pipelines, and jobs—and the generic item tools fill many lifecycle gaps. Some areas remain asymmetric (connections/gateways/capacities are mostly read/list-only, reports/semantic models lack direct create/update), but agents can often work around these via generic tools.
One MCP endpoint for Claude, GPT & Gemini: 100+ tools + no-code connectors + agent workers.
The OpenRouter for tools. One MCP connection gives any AI agent 254 hosted tools, pay per call.
Discover, plan and run tools across your connected business systems through one OAuth MCP endpoint
Zero-setup MCP gateway securely connecting AI to your tools with authentication and workflows