mcpField
E
entra-news

entra-news

Local
github.com/darrenjrobinsonSearch7Updated 8w ago

About entra-news

entra-news is an MCP server for searching the Entra.news archive of weekly Microsoft Entra coverage. It helps people researching Entra features and community tools find sourced excerpts, issues, and dates.

Key features

  1. 01Search the archive by keyword and phrase
  2. 02Optional semantic and hybrid search
  3. 03Retrieve full issues by number or date
  4. 04Filter issues by year or month
  5. 05Find mentioned tools and GitHub projects

Use cases

  1. 01Find past coverage of an Entra feature
  2. 02Locate an issue by publication date
  3. 03Discover community tools mentioned in the newsletter

MCP support

The MCP server exposes tools to search, retrieve, and browse Entra.news issues, and to find tools mentioned in the archive.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"darrenjrobinson-entra-news": {
"command": "npx",
"args": [
"-y",
"entra-news-mcp"
]
}
}
}
Run commandnpx -y entra-news-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmentra-news-mcp0.1.5

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

entra-news-mcp

A searchable knowledge MCP over Entra.news — Merill Fernando's curated weekly digest of Microsoft Entra news, features, and community tools.

npm npm downloads License: MIT


What is this?

Entra.news is a high-signal, curated newsletter covering Microsoft Entra (Azure AD) features, announcements, and community tools — published weekly since mid-2023.

This MCP server exposes the full historical archive as a natural language search interface. Ask questions and get sourced answers directly from past issues — including issue number, date, and canonical URL.

Zero per-user infrastructure. Users install an NPX package. That's it.


Quick Start

Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows):

{
  "mcpServers": {
    "entra-news-mcp": {
      "command": "npx",
      "args": ["entra-news-mcp"]
    }
  }
}

Restart Claude Desktop. The database (~15–20 MB) will be downloaded on first launch and cached in ~/.entra-news-mcp/.

Cursor / Copilot Studio / Any MCP Host
{
  "mcpServers": {
    "entra-news-mcp": {
      "command": "npx",
      "args": ["-y", "entra-news-mcp"]
    }
  }
}
VS Code

Add to your MCP config (workspace .vscode/mcp.json, or run MCP: Add Server from the Command Palette):

{
  "servers": {
    "entra-news-mcp": {
      "command": "npx",
      "args": ["-y", "entra-news-mcp"]
    }
  }
}
Semantic Search (Optional)

By default the server uses keyword and phrase matching over the SQLite index — no API key needed. For significantly better result quality, set your OpenAI API key to enable semantic + hybrid search:

{
  "mcpServers": {
    "entra-news-mcp": {
      "command": "npx",
      "args": ["entra-news-mcp"],
      "env": {
        "OPENAI_API_KEY": "sk-..."
      }
    }
  }
}

Local Database Cache

On first launch the server downloads the database (~15–20 MB) from GitHub Releases and caches it locally:

PlatformCache location
Windows%USERPROFILE%\.entra-news-mcp\
macOS / Linux~/.entra-news-mcp/

The server checks for a newer database release once per week. If you want to force an immediate re-download (e.g. after a new issue has been ingested), delete the cache folder and restart your MCP host:

Windows (PowerShell):

Remove-Item "$env:USERPROFILE\.entra-news-mcp" -Recurse -Force

macOS / Linux:

rm -rf ~/.entra-news-mcp

Available MCP Tools

ToolDescription
search_entra_newsSemantic + keyword hybrid search over all issues. Returns sourced excerpts.
get_issueRetrieve the full content of a specific issue by number or date.
list_issuesBrowse the archive with optional year/month filtering.
find_tool_mentionsDiscover community tools and GitHub projects mentioned in the archive.
Example queries
  • "What did Entra.news cover about Conditional Access in 2024?"
  • "Show me the issue from March 2025"
  • "What PowerShell tools for Entra have been mentioned?"
  • "Has there been coverage of Verified ID?"
  • "List all issues from 2024"

Architecture

Substack API (entra.news/api/v1/posts)
    │
    ▼
Node.js ingestion script  ←  OpenAI text-embedding-3-small
    │
    ▼
SQLite (chunks + embeddings, ~15–20 MB)
    │
    ▼
GitHub Release asset  ──→  NPX MCP Server
                               └─ Downloads DB on first run
                               └─ Checks for updates weekly
                               └─ In-memory vector similarity + keyword search

Cost: ~$0.01/week (embeddings on new issues only). Zero hosting.


Running the Ingestion Pipeline

Note: You only need to do this if you're maintaining your own fork or building the initial index. End users just run npx entra-news-mcp — the database is downloaded automatically.

Prerequisites
  • Node.js 22+
  • An OpenAI API key (text-embedding-3-small access)
Full ingest (first time)
# Set your API key
$env:OPENAI_API_KEY = "sk-..."

# Run the ingestion pipeline
./scripts/ingest.ps1

Or directly with Node.js:

export OPENAI_API_KEY=sk-...
npm install && npm run build
node dist/scripts/ingest.js
Incremental update (new issues only)
./scripts/ingest.ps1 -Incremental
node dist/scripts/ingest.js --incremental

The output database (entra-news.db) should then be uploaded as a GitHub Release asset — the GitHub Actions workflow handles this automatically on a weekly schedule.


Automated Weekly Updates

A GitHub Actions workflow (.github/workflows/weekly-update.yml) runs every Monday at 9am Sydney time (Sunday 23:00 UTC), shortly after each new Entra.news issue is published:

  1. Downloads the current database from GitHub Releases
  2. Runs the incremental ingestion pipeline
  3. Publishes the updated database as a new GitHub Release

Required secret: Add OPENAI_API_KEY to your repository secrets (Settings → Secrets).


Releasing (npm + MCP Registry)

Publishing is automated by .github/workflows/publish-mcp.yml, triggered by pushing a v* tag. Authentication is tokenless (OIDC) for both npm (Trusted Publishing) and the MCP Registry.

  1. Bump the version in package.json and server.json (both version fields) — the workflow fails if they don't match the tag
  2. Commit, then tag and push — deriving the tag from package.json so it always matches the workflow's version gate:
VERSION=$(node -p "require('./package.json').version")
git tag "v$VERSION"
git push origin main "v$VERSION"

The workflow builds, publishes to npm (with provenance), and publishes the new version to the MCP Registry.


Development

npm install
npm run build          # Compile TypeScript
npm start              # Run the MCP server
Project structure
src/
  index.ts             # Entry point
  server.ts            # MCP server + tool registration
  db/
    client.ts          # SQLite client — DB download/cache + search
  tools/
    search.ts          # search_entra_news tool
    get-issue.ts       # get_issue tool
    list-issues.ts     # list_issues tool
    find-tool-mentions.ts  # find_tool_mentions tool
  utils/
    embeddings.ts      # OpenAI embedding helper
scripts/
  ingest.ts            # Full ingestion pipeline (TypeScript)
  ingest.ps1           # PowerShell wrapper for ingestion
.github/workflows/
  weekly-update.yml    # Automated weekly update

Permissions & Content

The Entra.news content is © Merill Fernando & Joshua Fernando. This tool accesses the publicly available Substack API (not scraping) and is intended for personal/community use. Please reach out to [email protected] before any public deployment.


Author

Built by Darren Robinson.

Entra.news by Merill Fernando.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.