MCP Server · 21 tools · Works with any MCP client

Your AI assistant just got a news desk.

Claude, ChatGPT, and Cursor only know their training data. Agentic gives them real-time, personalized news intelligence — the stories that matter to your work, not everyone’s.

Works with:Claude DesktopChatGPTCursorVS CodeWindsurfAny MCP client

Set up in 30 seconds

One command. One config block. Done.

1Get your API key

2Add to your config

Paste into your Claude Desktop, Cursor, or VS Code MCP config:

{
  "mcpServers": {
    "agentic": {
      "command": "npx",
      "args": ["-y", "@agentic/mcp-server"],
      "env": {
        "AGENTIC_API_KEY": "ak_your_key_here"
      }
    }
  }
}

That’s it. Ask your AI assistant “What’s in my briefing today?”

Your AI assistant is blind to what happened today.

Claude and ChatGPT know their training data. They don’t know about this morning’s funding round, yesterday’s product launch, or the regulatory change that affects your roadmap.

Generic news MCP servers dump raw articles into your context window. No deduplication. No relevance scoring. No memory of what you’ve already seen.

Agentic is different. It’s not a news API — it’s a personal news analyst that learns what matters to you.

Clusters 10,000+ sources

Deduplicates and groups related articles

Agents monitor your topics

Custom watchers that learn from your feedback

AI-scored relevance

Every match validated by LLM for your specific domain

Stateful delivery

"Since last check" — never see the same story twice

AI briefs, not raw articles

Smart Brevity format: what happened, why it matters

Read-only, transparent

We read news. We never write, post, or act on your behalf.

21 tools. 5 categories.

Focused, not bloated. Every tool solves a daily workflow problem.

Briefing3 tools

Morning briefing, trending stories, deep dives — personalized to your agents.

get_briefingget_trendingget_story

Agents4 tools

Create monitoring agents for any topic. They learn from your feedback.

list_agentscreate_agentupdate_agentget_agent_stories

Search1 tools

Semantic search across 10,000+ sources. Not keyword matching — meaning matching.

search_stories

Feedback3 tools

Thumbs up/down trains your agents. Refinements fine-tune what matters.

submit_feedbacklist_refinementsadd_refinement

Collections3 tools

Save and organize stories. Build research libraries your AI can reference.

list_collectionscreate_collectionsave_to_collection

Ask your assistant anything.

Real examples of what Agentic enables.

You:What happened in AI infrastructure overnight?

via AI Infrastructure Watch

Your AI Infrastructure agent found 3 stories since your last check: 1. Anthropic open-sources tool-use framework — new approach to structured outputs that could simplify your agent pipelines. 4 sources confirm. 2. AWS announces MCP-native Lambda triggers — serverless MCP servers without container management. Early access program open. 3. Databricks acquires vector DB startup — signals consolidation in the embedding infrastructure space.

You:Create an agent to track LLM model releases

via New agent

Created agent LLM Model Releases. Scanning 10,000+ sources for model announcements, benchmark results, and capability updates. Your first briefing will be ready in ~2 minutes. Tip: Use `submit_feedback` on stories to teach it what's relevant to your work vs. general AI hype.

You:Search for recent MCP security vulnerabilities

via Semantic search

Found 5 stories matching MCP security vulnerabilities (last 7 days): 1. CVE-2025-6514 patched in MCP SDK 1.8.2 — the RCE via tool poisoning that affected servers using stdio transport. Upgrade immediately if you're on <1.8.0. 2. Invariant Labs releases MCP Scan 2.0 — automated vulnerability scanning for MCP server configs, now covers rug-pull detection.

Built for people who live in their AI tools.

AI Engineers

Track model releases, framework updates, and infrastructure shifts without leaving Claude or Cursor.

Technical Founders

Monitor competitors, funding rounds, and market signals — briefed by the same AI assistant you code with.

Research Analysts

Feed domain-specific intelligence into ChatGPT for synthesis. Ask follow-up questions against live data.

DevRel & Developer Advocates

Track ecosystem news, community sentiment, and competing SDKs. Your agents catch what Twitter misses.

Personal intelligence, not enterprise pricing.

Enterprise CI platforms charge $25,000+/year. Agentic starts at $9/month.

Monthly

$9/mo

14-day free trial

Annual (save 27%)

$79/yr

14-day free trial

Includes unlimited agents, AI briefs, semantic search, and full MCP access.

Stop asking your AI assistant to guess.

Give it real-time intelligence. 30 seconds to set up. 2 minutes to your first briefing.

No credit card required for 14-day trial.