azdo-onprem-mcp

azdo-onprem-mcp

Minimal MCP server for self-hosted Azure DevOps Server (on-premises). Enables interaction with work items, WIQL search, and project listing via REST APIs using PAT and basic auth.

Category
Visit Server

README

azdo-onprem-mcp

Minimal Model Context Protocol server for self-hosted Azure DevOps Server (on-premises). It uses the REST APIs with a PAT and Basic auth (curl -u ":$PAT"), same as Azure DevOps Services patterns but with your own base URL.

Not for dev.azure.com only—any DevOps Server reachable over HTTPS works if the APIs respond.

Requirements

  • Node.js 18+
  • A Personal Access Token with at least Work Items (Read) (and any other scopes you rely on)
  • VPN/network access to your server if required

Install

npm install -g azdo-onprem-mcp

Or run without a global install:

npx azdo-onprem-mcp

The package exposes the CLI binary azdo-onprem-mcp (see package.jsonbin).

Environment variables

Variable Required Description
AZURE_BASE_URL Yes Team project root URL, same as in the browser for that project, e.g. https://devops.example.com/Collection/MyProject (no trailing slash required).
AZURE_PAT Yes Personal Access Token. Sent as Basic auth with an empty username (Authorization: Basic base64(":PAT")).
AZURE_COLLECTION_URL No Overrides collection root for listProjects only. If unset, the collection URL is derived by dropping the last path segment of AZURE_BASE_URL (e.g. .../Collection/Project.../Collection).
AZURE_HTTP_TIMEOUT_MS No HTTP timeout in ms (default 120000).
AZURE_DEVOPS_API_VERSION No REST api-version query parameter (default 7.1). Try 7.0 or 6.0 if your server is older.
MCP_DEBUG No Set to 1 to log request URLs and timing on stderr (no secrets).

Do not commit secrets. Prefer Cursor/IDE env injection or your OS secret store.

Cursor (or any MCP client)

Example mcp.json entry using the published package (-y lets npx install or run without an interactive prompt, which MCP clients need):

{
  "mcpServers": {
    "azdo-onprem": {
      "command": "npx",
      "args": ["-y", "azdo-onprem-mcp"],
      "env": {
        "AZURE_BASE_URL": "https://devops.example.com/Collection/MyProject",
        "AZURE_PAT": "<your-pat>"
      }
    }
  }
}

Alternatives

  • Global install: "command": "azdo-onprem-mcp" with no args (or an empty args array), if npm install -g azdo-onprem-mcp put the binary on your PATH.
  • Local clone: after npm install and npm run build, use "command": "node" and "args": ["/absolute/path/to/azdo-onprem-mcp/dist/server.js"] (adjust the path to your machine).

Reload MCP / restart the editor after changing env.

Windows / npx: From v1.0.1, HTTP calls use Node’s built-in fetch only (the axios dependency was removed). That avoids a known failure where npx pulled a broken tree and mime-db could not load db.json. If you still see strange missing-module errors, run npm cache clean --force and try again.

Tools

Tool Description
getWorkItem GET/_apis/wit/workitems/{id}? $expand=all — returns the full REST body as MCP structuredContent and as JSON text. Empty field values are often omitted by Azure DevOps; unset fields are not listed. If the chat UI shows a short snippet, use the structured payload (or your client’s tool-result JSON view).
searchWorkItems WIQL search on System.Title (contains), then batch-fetch details — returns { id, title, state }[].
listProjects GET/_apis/projects at the collection URL (see AZURE_COLLECTION_URL / derivation above).

Verify with curl (optional)

Mac/Linux: use single quotes around the URL so $expand is not interpreted by the shell.

export AZURE_PAT='your-pat'
curl -sS -u ":$AZURE_PAT" \
  'https://your-host/Collection/MyProject/_apis/wit/workitems/12345?$expand=all&api-version=7.1'

Development

From a clone of this repository:

npm install
npm run build
npm start

Source lives in src/ (TypeScript); npm run build emits to dist/. With AZURE_BASE_URL and AZURE_PAT set, npm start runs the MCP server on stdio (stop with Ctrl+C). Point your MCP client at this command or use npx azdo-onprem-mcp after publishing.

Publish to npm

From the repo root (with npm login if needed):

npm run build
npm publish

prepack runs npm run build automatically, so dist/ is always fresh in the tarball. For a scoped package name (e.g. @your-scope/azdo-onprem-mcp), use npm publish --access public the first time.

License

MIT (see package.json"license").

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured