Newsdesk MCP for News-based Structured Intelligence

by NewsDeskDev

262 downloads
Not rated
GitHub

About

Newsdesk MCP for News-based Structured Intelligence is a hosted Model Context Protocol server that connects AI agents to NewsDesk's live structured news intelligence database. It converts raw news articles into a structured database of events, entities (17 types), relationships…

Details

Author
NewsDeskDev
Downloads
262
Categories
AI, Finance

- No installation required – hosted MCP server.
- 17 entity types including people, companies, countries, and more.
- 17 tools for discovery, entity, event, relationship, action, and graph queries.
- Semantic search and structured filters for events.
- Pathfinding and multi-hop graph pattern matching.
- OAuth 2.0 with Google, no API keys to manage.

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name Newsdesk MCP for News-based Structured Intelligence
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Create a .mcp.json file in your MCP client's working directory with the type http and URL https://app.newsdesk.dev/mcp. Restart the client; on first connection, a browser window opens for Google OAuth sign-in (a NewsDesk account is required). After signing in, you can start querying the agent with natural language questions about news entities, events, and relationships.

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "newsdesk mcp for news-based structured intelligence": {
            "newsdesk": {
                "type": "http",
                "url": "https://app.newsdesk.dev/mcp"
            }
        }
    }
}

McpServers

{
    "newsdesk": {
        "type": "http",
        "url": "https://app.newsdesk.dev/mcp"
    }
}

NewsDesk MCP Server

Connect AI agents to NewsDesk's live structured news intelligence database via the Model Context Protocol.

This is a hosted MCP server. No installation required — connect any MCP-compatible client to https://app.newsdesk.dev/mcp and authenticate with Google.

What is NewsDesk?

NewsDesk converts raw news articles into a structured database of events, entities (people, companies, countries — 17 types), relationships, and actions. The MCP server gives AI agents direct access to query this data using natural language.

Quick Start (2 minutes)

1. Create .mcp.json

Add this file to the directory where you run your MCP client:

{
  "mcpServers": {
    "newsdesk": {
      "type": "http",
      "url": "https://app.newsdesk.dev/mcp"
    }
  }
}

No API keys or tokens to manage.

2. Restart your client

Claude Code, Cursor, Windsurf, and other MCP clients pick up .mcp.json on startup.

3. Sign in

On first connection, your client opens a browser for Google sign-in. Use the same account you use for the NewsDesk dashboard. Your session is cached after the first sign-in.

4. Start querying

Ask your agent questions like:

- "What's happening with NVIDIA and China?"
- "How are Elon Musk and the SEC connected?"
- "Find all government agencies that took negative actions against tech companies"
- "Find the path between Donald Trump and Saudi Arabia"

The agent automatically selects the right tools to answer.

Tools (17)

| Category | Tools | What they do |
|----------|-------|-------------|
| Discovery | get_schema, get_stats | Data model reference, database statistics |
| Entity | search_entities, get_entity, get_entity_events | Find and inspect any of 17 entity types |
| Event | get_event, get_events_batch, search_events, similar_events | Semantic search, structured filters, similarity, bulk fetch |
| Relationship | get_entity_relationships, get_relationship, entity_co_mentions | Browse connections, pair details, co-mention strength |
| Action | get_entity_actions, get_actions_between, search_actions | Who did what to whom — sanctions, acquisitions, agreements |
| Graph | graph_query, find_paths | Multi-hop pattern matching, shortest paths between entities |

See TOOLS.md for full parameter documentation and examples.

Entity Types

| Code | Type | Example |
|------|------|---------|
| per | Person | per___Donald_Trump |
| stock | Public company | stock___NVIDIA |
| priv | Private company | priv___SpaceX |
| subs | Subsidiary | subs___YouTube |
| cnt | Country | cnt___China |
| govactor | Government agency | govactor___United_States===Federal_Reserve |
| cbnk | Central bank | cbnk___Japan===Bank_of_Japan |
| polparty | Political party | polparty___United_States===Republican_Party |
| curr | Currency | curr___US_Dollar |
| cmdt | Commodity | cmdt___Crude_Oil |
| crypto | Cryptocurrency | crypto___Bitcoin |
| index | Market index | index___S&P_500 |
| exch | Stock exchange | exch___NYSE |
| alliance | International alliance | alliance___NATO |
| ngo | NGO | ngo___Amnesty_International |
| mil | Military | mil___United_States===US_Army |
| loc | Location | loc___Silicon_Valley |

Entity reference format: type___Name (triple underscore). Parent-child: type___Parent===Child.

Example Workflows

Entity lookup
> "Tell me about Tesla"

Agent calls: search_entities("Tesla")get_entity("stock___Tesla_Inc.")get_entity_events("stock___Tesla_Inc.")

Relationship exploration
> "How are the US and China connected?"

Agent calls: search_entities("United States")get_entity_relationships("cnt___United_States")get_relationship("cnt___United_States", "cnt___China")

Semantic search
> "What events are related to AI regulation in Europe?"

Agent calls: search_events(query="AI regulation Europe")

Complex graph pattern
> "Find all CEOs of companies that have relationships with China"

Agent calls: graph_query({nodes: {CHINA: anchor, COMPANIES: intermediate, CEOS: result}, edges: [...]})

Pathfinding
> "How is Elon Musk connected to the Federal Reserve?"

Agent calls: find_paths("per___Elon_Musk", "govactor___United_States===Federal_Reserve")

Supported Clients

- Claude Code (CLI, desktop, web, IDE extensions)
- Cursor
- Windsurf
- Any client supporting the MCP specification

Authentication

OAuth 2.0 via Google. The MCP client handles the flow automatically — a browser window opens on first connection, and the session is cached afterward. No API keys to manage.

Requires a NewsDesk account. Request access if you don't have one.

Links

- Full tutorial
- NewsDesk Dashboard
- API Documentation
- Pricing

License

MIT — see LICENSE.

No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.