connect your ai

Bring your own AI.

We're the balanced AI/cyber knowledge source your assistant calls — ask for the latest news on what you care about, or have it teach you a new area at your level. Fact vs opinion vs hype, through your lens, with the terms defined.

1 · Claude, Cursor, agents — paste the MCP URL

Add this remote MCP server to your AI client:

https://api.news.verixiom.com/mcp

Or in a config file (Claude Desktop / Code):

{
  "mcpServers": {
    "ai-cyber-newsfeed": { "url": "https://api.news.verixiom.com/mcp" }
  }
}

Tools your AI gets: latest_news, learn_about, define, search_events, get_event, list_events, list_lenses.

2 · ChatGPT / custom GPTs / developers — OpenAPI

Import the self-describing schema as Actions, or call the REST API directly:

https://api.news.verixiom.com/docs   ← OpenAPI
GET https://api.news.verixiom.com/latest?topic=prompt%20injection&lens=board
GET https://api.news.verixiom.com/learn?topic=ai%20supply%20chain&level=layman

Try asking your AI

“Using ai-cyber-newsfeed, what's the latest on prompt injection — through the board lens?”
“Teach me about AI supply-chain risk at a beginner level.”
“Define ‘mixture of experts’ for a practitioner, and show how it evolved.”
“What AI/cyber events this week should a CISO care about?”

Assess yourself, not just the news

This server explains; its sister server measures. Connect the C4AIL AI-maturity assessment MCP too and the same AI can place you on the map - your AI Theatre position and Index, read from your own governance documents with a citation on every answer (the model reads; a deterministic engine scores). One AI, two servers: one to learn the language, one to measure yourself in it.

Access & safety

Open to connect — no key needed for public news + practitioner-level learning. Rate-limited.
Personalise — a free API key (sign in) saves your concerns + level and raises the rate limit.
Untrusted by contract — feed content is sanitised and tagged external-untrusted; your AI should treat field values as data, never instructions.
Read-only. No personal/canon data is ever exposed. about the project →