Skip to main content

TOMPo — Power BI & Fabric Lineage Intelligence MCP Server. Trace lineage from semantic models to reports to visuals. Impact analysis for every column and measure.

Project description

TOMPo MCP — Power BI & Fabric Lineage Intelligence

Trace lineage from semantic models → tables → reports → pages → visuals → columns/measures.

An MCP (Model Context Protocol) server that brings Power BI lineage intelligence directly into your AI assistant — GitHub Copilot, Claude Desktop, Cursor, or any MCP-compatible client.

What it does

  • Full Lineage: See exactly which columns and measures appear in which visuals, across all reports bound to a semantic model
  • Impact Analysis: "What breaks if I rename Employee.StartDate?" — instantly shows every affected visual
  • Interactive Visualization: Export a self-contained D3 tree (HTML file) with expand/collapse, zoom, and search
  • Zero Infrastructure: Runs locally on your machine using your own Azure identity. No App Service, no Docker, no backend.

Quick Start

1. Install

pip install git+https://github.com/microsoft/tompo-mcp.git

Or clone and install locally:

git clone https://github.com/microsoft/tompo-mcp.git
cd tompo-mcp
pip install -e .

2. Login to Azure

az login

3. Add to VS Code

Add to your VS Code settings.json (Ctrl+Shift+P → "Preferences: Open User Settings (JSON)"):

{
  "mcp": {
    "servers": {
      "tompo": {
        "command": "python",
        "args": ["-m", "tompo_mcp"]
      }
    }
  }
}

4. Use in Copilot Chat

Open GitHub Copilot Chat and start asking:

> List my Fabric workspaces
> Generate lineage for dataset abc-123 in workspace xyz-456
> What visuals use the Employee.Department column?
> Export the lineage as an interactive HTML file

MCP Tools

Tool Description
list_workspaces List all Fabric/Power BI workspaces you have access to
generate_lineage Full lineage: Model → Tables → Reports → Pages → Visuals → Fields
impact_analysis Find all visuals where a specific column or measure is used
describe_semantic_model Detailed metadata: tables, columns, measures, relationships, roles
export_lineage_html Generate interactive D3 visualization as a self-contained HTML file

How It Works

You type in Copilot Chat
        │
        ▼
Copilot calls TOMPo MCP tools
        │
        ▼
TOMPo runs locally on your machine:
  → Uses your az login identity
  → Calls Fabric REST APIs (getDefinition, Scanner, DAX)
  → Parses model + report definitions
  → Builds lineage tree
  → Returns data to Copilot
        │
        ▼
Copilot shows the lineage tree / impact table
(or opens interactive HTML in your browser)

Authentication

TOMPo uses DefaultAzureCredential which automatically picks up:

  1. Azure CLI (az login) — most common for developers
  2. VS Code Azure Account — if you're signed into the Azure extension
  3. Environment variables — for CI/CD pipelines
  4. Managed Identity — for Azure-hosted scenarios

You need access to the Fabric workspaces you want to analyze. No extra app registrations or service principals required.

Metadata Extraction (3-tier fallback)

  1. Fabric getDefinition API — returns full model.bim / TMDL / PBIR definitions
  2. Admin Scanner API — fallback if getDefinition fails (requires admin permissions)
  3. DAX executeQueries — last resort using INFO.TABLES(), INFO.COLUMNS(), etc.

If a sensitivity label blocks access, TOMPo temporarily downgrades to "General", extracts metadata, then restores the original label.

Interactive Visualization

The export_lineage_html tool generates a single HTML file with:

  • D3 horizontal tree with expand/collapse nodes
  • Color-coded by type (model, table, report, page, visual, column, measure)
  • Impact Analysis tab with searchable data grid
  • Semantic Model tab with tables, columns, measures, relationships
  • Zoom, pan, fullscreen — all interactive
  • Works offline — all JavaScript and CSS inlined, no server needed

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "tompo": {
      "command": "python",
      "args": ["-m", "tompo_mcp"]
    }
  }
}

Development

git clone https://github.com/microsoft/tompo-mcp.git
cd tompo-mcp
pip install -e ".[dev]"
python -m pytest tests/

Requirements

  • Python 3.10+
  • Azure CLI (az login) or any Azure credential
  • Access to Fabric/Power BI workspaces

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

tompo_mcp-0.1.0.tar.gz (28.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

tompo_mcp-0.1.0-py3-none-any.whl (29.4 kB view details)

Uploaded Python 3

File details

Details for the file tompo_mcp-0.1.0.tar.gz.

File metadata

  • Download URL: tompo_mcp-0.1.0.tar.gz
  • Upload date:
  • Size: 28.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.10

File hashes

Hashes for tompo_mcp-0.1.0.tar.gz
Algorithm Hash digest
SHA256 0fd8ef17c16d90ee57aa0c9318323441c12672e3fcebd67effd5695284462efe
MD5 8cd6f8133d12e1ece95665498696b19c
BLAKE2b-256 1d57c55f42cd8dc408b51d2560277a34a32c3dbee8a7f6d4ab26032adb7d95df

See more details on using hashes here.

File details

Details for the file tompo_mcp-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: tompo_mcp-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 29.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.10

File hashes

Hashes for tompo_mcp-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 8343636b8feae273db799636294150012d1d8c206cf311ae587edfe5dabd0628
MD5 b318e84cfcbcc4e1ca6c163b8690f951
BLAKE2b-256 ad39a3a7aa9584dc906cc6c9977267fb0c3928c7a5ca7b25d4ffd857b7cfae68

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page