T&B Newswire — MCP Connector

Install in Claude (custom connector)

1. Open claude.ai, then Settings → Connectors → Add custom connector.
2. Paste the URL: https://techandbusiness.org/mcp
3. Leave authentication blank. Save.
4. In a new chat, enable the connector and try:

   Use the Tech & Business Newswire connector to list five recent AI stories with canonical URLs.

Claude will return results with full https://techandbusiness.org/newswire/{id} URLs, so it can cite stories directly back to the source.

Install in Claude Desktop (via mcp-remote)

Add the following block to your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "tb-newswire": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://techandbusiness.org/mcp"]
    }
  }
}

Restart Claude Desktop after saving.

Tools

Ten read-only tools. Every response includes canonical URLs. Newswire tools (seven) are listed first; article tools (three) follow.

list_recent_headlines

Recently published stories. Arguments: days (1-30, default 7), limit (1-100, default 25), coverage (optional: ai, tech-biz, startups, policy, cybersecurity, infrastructure), tier (1 major or 2 notable).

get_story

One story by Sanity _id (matches the slug on techandbusiness.org/newswire/{id}). Returns clean markdown with headline, body, sources, and canonical URL.

search_stories

Full-text search across headlines and bodies. Arguments: query (required), days (1-90, default 30), limit (1-25, default 10), coverage.

get_daily_digest

The seven top stories from a given day's sealed edition. Arguments: date (YYYY-MM-DD, optional; defaults to the most recent sealed edition).

list_categories

The six coverage areas plus 30-day story counts.

get_archive_window

All stories with publish dates in [from_date, to_date] inclusive, America/New_York. Arguments: from_date, to_date (both YYYY-MM-DD, required), coverage, limit (1-100, default 50).

get_weekly_brief

Returns a consequence-ranked and same-event-deduplicated rolling brief for recurring presentations, including editorial scores, score reasons, dominant houses, recurring entities, and source URLs.

list_articles

Long-form articles (evergreen editorial, distinct from the Newswire). Arguments: limit (1-50, default 20). Returns canonical URLs, markdown URLs, titles, excerpts, authors, publish dates.

get_article

One article by slug (matches the path at techandbusiness.org/articles/{slug}). Returns clean markdown with title, byline, excerpt, body, and canonical URL.

search_articles

Full-text search across article titles and bodies. Arguments: query (required), limit (1-25, default 10). No date window — articles are evergreen.

Related surfaces

• llms.txt — publication index for LLMs.
• llms-full.txt — last 60 days of published stories as one plain-text corpus, regenerated nightly.
• llms-articles.txt — all published long-form articles as one markdown corpus, regenerated nightly.
• /.well-known/mcp.json — MCP discovery card.
• /.well-known/agent-skills.json — skills index.
• Any /newswire/{id} URL responds with text/markdown when requested via Accept: text/markdown or ?format=md.

Privacy & abuse

No authentication, no per-user tracking. Standard Cloudflare access logs apply. The server is read-only and exposes only published, non-archived stories. If you need higher rate limits or a dedicated channel, write to [email protected].