M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
F

Fabric RTI MCP

Official
KQL queries, Eventstream management, and real-time alerts for Microsoft Fabric Real-Time Intelligence
How to install ↓ Compare GitHub source repository ↗
Category
Other #38 of 230
Stars
★ 131 Very popular
Transport
stdio (local process) · Streamable HTTP
Runtime
Python 3.10+ · Hosted (no local runtime)
Credentials
API key / credential required
License
MIT
Last commit
Tools
39
65FMRS · C

This is an official Microsoft MCP server that connects AI agents to Fabric RTI, offering a wide set of tools including KQL queries, event stream management, and map operations. It provides a natural language interface and is well-documented. However, it is in preview and requires careful permission management to use securely.

Strongest · Documentation 15/20 Weakest · Reliability 10/20

Reliability
10/20
Security and permissions
13/20
Maintenance
15/20
Documentation
15/20
Setup experience
12/20
Why each score
Reliability 10/20
The repository contains tests for the server's key paths (test files exist but not inspected), and CI workflows (specific files not provided but the repository is active on GitHub). The manifest and README show a structured list of tools, but static review cannot verify tools fully match behavior. No evidence of server failing to initialize, so 10 is appropriate.
Security and permissions 13/20
Uses Azure Identity for credential handling, no token storage. HTTP mode has bearer token screening and OBO support. Tool-level restriction via RBAC. No confirmation defaults for dangerous delete, deploy, or payment operations observed, but README mentions safeguards for destructive actions. No red flags found. Security score 13 because confirmation mechanisms not standardized in deployment, HTTP mode requires explicit security boundary.
Maintenance 15/20
Active repository with recent commits, and the 15 reflects ongoing maintenance. LICENSE present, clear Microsoft CLA and contributing guidelines. This score is based on repository activity without verified commit dates, but repository status shows not archived.
Documentation 15/20
README provides extensive documentation covering installation, configuration, authentication, environment variables, example prompts, and debugging. No explicit troubleshooting section, but documentation is generally detailed. Score 15 reflects minor gaps like lack of dedicated support channel or clearer security guidance.
Setup experience 12/20
Multiple installation methods provided (uvx, pip, source). VS Code configuration examples included. Steps are clear, but dependency on uv may be considered complex for some users. Score 12 because execution not verified and setup requires multiple steps.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account dataConnects to a database

Best for

  • Teams using Microsoft Fabric RTI who want to interact with data via AI agents using natural language.
  • Data engineers needing to quickly analyze streaming data and generate KQL queries automatically.
  • Developers looking to automate real-time data monitoring and alerts, such as Activator triggers and Eventstream management.
  • Analysts needing to visualize geospatial data and interact with Map items.

Not for

  • Production deployments seeking a stable, GA MCP server (currently in Public Preview).
  • Users unfamiliar with Fabric RTI or KQL, as the server relies on existing knowledge to use effectively.
  • Organizations that do not wish to expose their data or workspaces to LLMs or third-party clients.
  • Use cases requiring advanced security for HTTP mode without additional configuration.

Required permissions

  • Access Microsoft Fabric workspaces, including Eventhouse, Eventstreams, Activator, and Map items.
  • Execute KQL queries and management commands, including ingesting data and modifying schemas.
  • Create, update, and delete Eventstreams and Map items.
  • Create Activator triggers and configure email or Teams notifications.
  • Authenticate using Azure Identity, which automatically uses environment credentials, Azure CLI, PowerShell, etc.

Risks and side effects

  • Public Preview: implementation may change significantly prior to General Availability.
  • MCP clients can perform destructive operations (e.g., deleting Eventstreams or Maps) if permissions are too broad.
  • Sharing data with third-party LLMs or clients may fall outside Fabric's compliance boundary; review carefully.
  • HTTP mode may expose endpoints beyond localhost; ensure a security boundary (e.g., Entra authentication) is placed in front.
  • Some tools, like kusto_get_shots, require additional configuration (e.g., embedding endpoints and table).

Setup

Before you start

Runtime：Python 3.10+ · Hosted (no local runtime)

Other optional settings (27)
KUSTO_SERVICE_URI optional Default Kusto cluster URI, e.g. https://help.kusto.windows.net/, from your Azure Data Explorer cluster page.
KUSTO_SERVICE_DEFAULT_DB optional Default database name for Kusto queries, e.g. Samples.
FABRIC_API_BASE optional Base URL for the Microsoft Fabric API, default https://api.fabric.microsoft.com/v1.
AZ_OPENAI_EMBEDDING_ENDPOINT optional Azure OpenAI endpoint used by kusto_get_shots for AOAI embeddings, including deployment name and api-version.
KUSTO_KNOWN_SERVICES optional JSON array of preconfigured Kusto services.
KUSTO_EAGER_CONNECT optional Whether to eagerly connect to the default service on startup; default false.
KUSTO_ALLOW_UNKNOWN_SERVICES optional Security setting allowing connections to services not in the known services list.
KUSTO_SHOTS_TABLE optional Enables kusto_get_shots and sets its default shots table.
KUSTO_SHOTS_EMBEDDING_METHOD optional Default embedding method for kusto_get_shots (slm or aoai).
KUSTO_SHOTS_SLM_MODEL optional Default SLM model for kusto_get_shots, default harrier-v1-270m.
FABRIC_BASE_URL optional Base URL for the Microsoft Fabric web interface.
FABRIC_RTI_ALLOWED_TOOLS optional Comma-separated service names or full tool names to expose.
FABRIC_RTI_KUSTO_DEEPLINK_STYLE optional Override auto-detection of deeplink style (adx or fabric).
FABRIC_RTI_TRANSPORT optional Server transport mode, default stdio, can be http.
FABRIC_RTI_HTTP_HOST optional HTTP server host address, default 127.0.0.1.
FABRIC_RTI_HTTP_PORT optional HTTP server port, default 3000.
FABRIC_RTI_HTTP_PATH optional HTTP path for the MCP endpoint, default /mcp.
FABRIC_RTI_STATELESS_HTTP optional Whether to use stateless HTTP mode.
FABRIC_RTI_HTTP_ALLOW_MI optional Allow requests without a bearer token to use Managed Identity; only behind a trusted boundary.
FABRIC_RTI_HTTP_DEBUG_MODE optional Local HTTP testing mode with local credentials and permissive CORS; not for production.
FABRIC_RTI_HTTP_ALLOWED_HOSTS optional Host allow-list for HTTP DNS-rebinding protection.
FABRIC_RTI_HTTP_ALLOWED_ORIGINS optional Origin allow-list for HTTP DNS-rebinding protection.
FABRIC_RTI_CORS_ORIGINS optional CORS allowed origins; loopback origins if unset.
FABRIC_RTI_KUSTO_KNOWN_SERVICES_PROBE optional Probe policy for kusto_known_services (auto, always, never).
USE_OBO_FLOW optional Enable the OBO token exchange flow.
FABRIC_RTI_MCP_AZURE_TENANT_ID optional Azure AD tenant ID.
FABRIC_RTI_MCP_USER_MANAGED_IDENTITY_CLIENT_ID optional User managed identity client ID.

Install from PyPI:

  1. Ensure UV and VS Code are installed.
  2. Open VS Code and open the command palette (Ctrl+Shift+P).
  3. Run MCP: Add Server, select "Install from Pip".
  4. Enter the package name microsoft-fabric-rti-mcp and follow the prompts.
  5. Add configuration to settings.json or mcp.json, e.g., using uvx command.

Alternatively, if GitHub Copilot CLI is enabled, run /mcp add in a session and select the package.

MCP config · JSON 
{
  "mcpServers": {
    "fabric-rti-mcp": {
      "command": "uvx",
      "args": [
        "microsoft-fabric-rti-mcp"
      ],
      "env": {
        "KUSTO_SERVICE_URI": "https://help.kusto.windows.net/",
        "KUSTO_SERVICE_DEFAULT_DB": "Samples",
        "FABRIC_API_BASE": "https://api.fabric.microsoft.com/v1"
      }
    }
  }
}

Works in clients that read an "mcpServers" key. For clients such as VS Code that use a different key, the configurator below converts it.

.vscode/mcp.json
{
  "servers": {
    "fabric-rti-mcp": {
      "command": "uvx",
      "args": [
        "microsoft-fabric-rti-mcp"
      ],
      "env": {
        "KUSTO_SERVICE_URI": "https://help.kusto.windows.net/",
        "KUSTO_SERVICE_DEFAULT_DB": "Samples",
        "FABRIC_API_BASE": "https://api.fabric.microsoft.com/v1"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal
claude mcp add fabric-rti-mcp -e KUSTO_SERVICE_URI=https://help.kusto.windows.net/ -e KUSTO_SERVICE_DEFAULT_DB=Samples -e FABRIC_API_BASE=https://api.fabric.microsoft.com/v1 -- uvx microsoft-fabric-rti-mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After installing with uvx microsoft-fabric-rti-mcp, restart the MCP client and confirm Fabric RTI tools like kusto_query and eventstream_list appear in the tool list; try the prompt 'List my Kusto tables' — a successful answer proves the connection.

Troubleshooting

  1. Check authentication: ensure you are logged in via Azure CLI, Visual Studio, Azure PowerShell, or browser, and have appropriate permissions.
  2. Check environment variables: verify KUSTO_SERVICE_URI and KUSTO_SERVICE_DEFAULT_DB are set correctly.
  3. Check permissions: assign the user the required Fabric RBAC roles to perform queries, creates, or deletes.
  4. Check network: ensure the client can reach Fabric and Azure Data Explorer endpoints.
  5. For remote deployments, verify HTTP mode configuration (e.g., FABRIC_RTI_HTTP_PORT) and authentication tokens.

Things to try

Once connected, you can ask your AI assistant things like:

  • Get databases in my Eventhouse
  • Sample 10 rows from the StormEvents table
  • List all Eventstreams in my workspace
  • Using the StormEvents table, notify me via email when there is a flood in Illinois

Tools 39

kusto_known_services read-only
List all available Kusto services configured in the MCP.
kusto_query read-only
Execute KQL queries on the specified database.
kusto_command writes
Execute Kusto management commands (e.g., .show, .create, .alter, .drop).
kusto_list_entities read-only
List entities (databases, tables, external tables, materialized views, functions, graphs) in a cluster or database.
kusto_describe_database read-only
Get schema information for all entities in a database.
kusto_describe_database_entity read-only
Get detailed schema for a specific entity (table, external table, materialized view, function, or graph).
kusto_graph_query read-only
Execute graph queries using snapshots or transient graphs.
kusto_sample_entity read-only
Retrieve sample records from a table, external table, materialized view, or function.
Show 31 more tools
kusto_ingest_inline_into_table writes
Ingest inline CSV data into a specified table.
kusto_get_shots read-only
Find semantically similar saved KQL queries using local SLM or Azure OpenAI embeddings.
kusto_deeplink_from_query read-only
Generate a deeplink URL to open a KQL query in Azure Data Explorer Web Explorer or Microsoft Fabric query workbench.
kusto_show_queryplan read-only
Retrieve the execution plan for a KQL query without running it. Returns planning stats, logical operator tree, and execution hints.
kusto_diagnostics read-only
Run a best-effort suite of cluster diagnostic commands and return a unified summary.
eventstream_list read-only
List all Eventstreams in your Fabric workspace.
eventstream_get read-only
Get detailed information about a specific Eventstream.
eventstream_get_definition read-only
Retrieve complete JSON definition of an Eventstream.
eventstream_create writes
Create new Eventstreams with custom configuration.
eventstream_update writes
Modify existing Eventstream settings and destinations.
eventstream_delete destructive
Remove Eventstreams and associated resources.
eventstream_start_definition read-only
Start a definition-building session for an Eventstream.
eventstream_get_current_definition read-only
Get the current in-progress Eventstream definition.
eventstream_clear_definition writes
Clear the current Eventstream definition being built.
eventstream_add_sample_data_source writes
Add a sample data source to the Eventstream definition.
eventstream_add_custom_endpoint_source writes
Add a custom endpoint source to the Eventstream definition.
eventstream_add_derived_stream writes
Add a derived stream to the Eventstream definition.
eventstream_add_eventhouse_destination writes
Add an Eventhouse destination to the Eventstream definition.
eventstream_add_custom_endpoint_destination writes
Add a custom endpoint destination to the Eventstream definition.
eventstream_validate_definition read-only
Validate an Eventstream definition.
eventstream_create_from_definition writes
Create an Eventstream from a definition.
eventstream_list_available_components read-only
List components that can be used in an Eventstream.
activator_list_artifacts read-only
List all Activator artifacts in a Fabric workspace.
activator_create_trigger writes
Create new Activator triggers with KQL source monitoring and email/Teams alerts.
map_list read-only
List all Map items in your Fabric workspace.
map_get read-only
Get detailed information about a specific Map item.
map_get_definition read-only
Retrieve the full JSON definition of a Map item.
map_create writes
Create a new Map item from a provided configuration.
map_update_definition writes
Replace the full JSON definition of an existing Map item.
map_update writes
Partially update properties of an existing Map item.
map_delete destructive
Delete a Map item and its associated configuration.

Use cases

Query Eventhouse using natural language and analyze trends.
Manage Eventstreams: list, create, update, or delete real-time data streams.
Create Activator triggers to get email or Teams alerts when conditions are met or anomalies occur.
Create and manage Map items from JSON configurations to visualize geospatial data.

Supported clients

VS Code
GitHub Copilot

Listed from the project's documentation, not tested by this site.

Overview

The Fabric RTI MCP server is a comprehensive Model Context Protocol (MCP) server implementation for Microsoft Fabric Real-Time Intelligence (RTI). It enables AI agents to interact with Fabric RTI services by providing tools that allow for KQL queries against Eventhouse and Azure Data Explorer, management of Eventstreams, creation of Activator triggers, and visualization/manipulation of Map items. The server supports natural language to KQL conversion, secure authentication via Azure Identity, and real-time data access. It is currently in Public Preview.

Similar servers

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 654 · Tools 11 Compare with this →
Compare with other MCP servers

Source revision 3c765f92ca50 Data synced 2026-10-12 Read the FMRS scoring method