Skip to main content
Glama

PBIXRay MCP Server V2

INSTALL.md2.94 kB
# INSTALL.md — PBIXRay MCP Server (Windows) This guide gets you set up quickly on Windows with PowerShell. For enterprise/team rollout, you can adapt these steps to Intune/SCCM/DSC. ## Prerequisites - Windows 10/11 (64-bit) - .NET Framework 4.7.2+ (typically preinstalled) - Power BI Desktop (current recommended) - Claude Desktop - ~200 MB disk space Optional: Admin rights are NOT required; install to a user-writable path. ## 1) Get the files - Obtain the packaged folder (or clone/extract) and place it somewhere stable. Recommended path: - `C:\Tools\pbixray-mcp-server` Avoid paths with spaces or special characters. ## 2) Verify basics (optional but recommended) Open Windows PowerShell and run from the project folder: ```powershell cd "C:\Tools\pbixray-mcp-server" ./scripts/test_connection.ps1 ``` This lists available tools and validates the local environment. ## 3) Configure Claude Desktop (or ChatGPT) From the project folder, run: ```powershell ./scripts/install_to_claude.ps1 ``` Then fully restart Claude Desktop (quit the app from the tray/Task Manager and reopen). The script adds an entry to your Claude Desktop config pointing to: - Command: `venv\Scripts\python.exe` - Args: `src\pbixray_server_enhanced.py` For ChatGPT desktop (with MCP support), run: ```powershell ./scripts/install_to_chatgpt.ps1 ``` Then restart the ChatGPT app. ## 4) Connect and test 1. Open Power BI Desktop and load a .pbix file 2. Wait 10–15 seconds for the model to finish loading 3. In Claude, say: “Detect my Power BI Desktop instances” 4. Then: “Connect to instance 0” 5. Ask: “What tables are in this model?” If the server responds with your model’s tables, you’re good to go. ## Updating - Safe update: back up your current folder, then replace with the new package. - If you change the install path, rerun `./scripts/install_to_claude.ps1` so Claude points to the new location. - To refresh Python packages (if needed): ```powershell ./venv/Scripts/pip.exe install --upgrade -r requirements.txt ``` ## Uninstalling / Cleanup 1. Close Claude Desktop completely 2. Remove the Claude Desktop MCP entry (or rerun `install_to_claude.ps1 -Remove` if supported) 3. Delete the `pbixray-mcp-server` folder Optionally remove logs under `logs/`. ## Troubleshooting quick tips - “No instances detected”: ensure Power BI Desktop is open with a .pbix file and wait a few seconds - “Not connected”: detect → connect → then run queries - Claude doesn’t see the server: rerun `install_to_claude.ps1`, verify `%APPDATA%\Claude\claude_desktop_config.json`, and restart Claude - Slow queries: use `TOPN()` to limit rows; check SE vs FE via performance analysis ## Team distribution (optional) Create a ready-to-share zip: ```powershell cd "C:\Tools\pbixray-mcp-server" ./scripts/package_for_distribution.ps1 ``` Share the resulting archive with users and point them to this INSTALL.md.

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/bibiibjorn-Finvision/MCP-Server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server