Causari

Open-source MCP server

Wikipedia for AI agents.

Give Claude Code, Cursor, and any MCP-compatible agent structured causal knowledge — causes, effects, confidence scores, and evidence on every link. Open data. Install in 60 seconds.

Terminal

$ npx @causari/mcp-server

# or add to .mcp.json:

{

"causari": {

"command": "npx",

"args": ["@causari/mcp-server"]

}

}

5 tools for your agent

Structured causal knowledge, not guesswork.

01 · causal_chain

Trace cause and effect.

"What caused the AI winter?" → Perceptron limitations → DARPA funding cuts → connectionism backlash. Every link has confidence scores and cited evidence.

02 · historical_resonance

Find patterns that rhyme.

"My startup is building an open protocol competing with a closed platform." → Matches: TCP/IP vs OSI, Linux vs Unix, Web vs AOL. Predictive value + historical exemplars returned.

03 · query_events

Search by domain, time, impact.

"High-impact technology events between 1990-2010" → World Wide Web, Linux, Google, iPhone, Bitcoin. Structured JSON. Filter by domain, year, impact score.

See the data visually

Same graph, beautiful canvas.

Powflow renders the same causal knowledge graph your agent queries. Drag, zoom, click any node to trace its causes and effects.

Powflow · live Full screen ↗

↑ Drag · scroll to zoom · click any node

Give your agent a memory.

Free and open-source. Install the MCP server, or join the waitlist for the hosted Pro tier — zero config, always-fresh data, priority latency.

Embed Powflow anywhere

One line. Drop into Notion, Substack, your blog.

Paste this snippet into any page that allows raw HTML. The demo runs standalone — no account required for visitors.

<iframe src="https://causari.ai/embed/canvas.html" width="100%" height="640" style="border:0;border-radius:14px;background:#010105" loading="lazy" sandbox="allow-scripts allow-same-origin allow-popups allow-forms" title="Powflow by Causari"></iframe>

Copied to clipboard