NodusAI MCP Server

AI-Powered Signals for Prediction Markets — accessible to any AI agent via MCP.

AI agents connect to this server to get Oracle signals for Polymarket and Kalshi prediction markets. Signals are generated by Gemini 2.5 Flash with real-time web grounding.

How it works

Agent → nodusai.app → connect wallet → pay $1 USDC → get session token
                                                              ↓
Agent → MCP Server (nodus_get_signal) → nodusai.app/api/prediction → signal

1. Visit nodusai.app and connect your wallet
2. Paste a Polymarket or Kalshi market URL
3. (Optional) Add your desired outcome (YES / NO)
4. Pay $1 USDC — confirmed on-chain
5. Get a session token good for 3 queries
6. Use the session token with nodus_get_signal in any MCP client

Payment model

• Cost: $1 USDC = 3 Oracle signal queries
• Networks: Base, Ethereum, Avalanche (any EVM chain)
• Token: USDC
• Non-custodial: payments go directly on-chain via nodusai.app
• Session: one payment = one session token = 3 queries (24h validity)

Available tools

| Tool | Description | |------|-------------| | nodus_pricing | View pricing and how to get a session token | | nodus_get_signal | Get an Oracle signal using your session token | | nodus_verify_signal | Audit grounding sources of a past signal | | nodus_query_history | Your recent query history | | nodus_admin_stats | Platform-wide stats (admin) | | nodus_admin_queries | Full query registry dump (admin) |

Signal format

Every Oracle response follows NodusAI's structured schema:

{
  "market_name": "Will the Fed cut rates in June 2026?",
  "predicted_outcome": "YES",
  "probability": 0.73,
  "confidence_score": "HIGH",
  "key_reasoning": "Recent FOMC minutes and inflation data suggest...",
  "grounding_sources": [
    { "title": "Reuters: Fed signals rate path", "url": "https://..." },
    { "title": "AP: CPI data June 2026", "url": "https://..." }
  ]
}

Deploy in 5 minutes

Option 1 — Railway (recommended)

1. Fork this repo on GitHub
2. Go to railway.app → New Project → Deploy from GitHub repo
3. Select your fork
4. Add environment variable: NODUSAI_API_BASE = https://nodusai.app
5. Railway auto-detects railway.json and deploys
6. Copy your Railway URL

Option 2 — Render (free tier)

1. Fork this repo
2. Go to render.com → New Web Service → connect your fork
3. Set Build command: npm install and Start command: node src/server-http.js
4. Add env var: NODUSAI_API_BASE=https://nodusai.app

Option 3 — Fly.io

fly launch --name nodusai-mcp
fly secrets set NODUSAI_API_BASE=https://nodusai.app
fly deploy

Connect AI agents

Claude Desktop

File: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "nodusai": {
      "url": "https://nodusai-mcp-production.up.railway.app/sse"
    }
  }
}

Cursor

File: ~/.cursor/mcp.json

{
  "mcpServers": {
    "nodusai": {
      "url": "https://nodusai-mcp-production.up.railway.app/sse",
      "transport": "sse"
    }
  }
}

Windsurf

File: ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "nodusai": {
      "serverUrl": "https://nodusai-mcp-production.up.railway.app/sse"
    }
  }
}

Claude Code (CLI)

claude mcp add --transport sse nodusai https://nodusai-mcp-production.up.railway.app/sse

Custom JS agent

import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { SSEClientTransport } from "@modelcontextprotocol/sdk/client/sse.js";

const client = new Client({ name: "my-agent", version: "1.0.0" }, { capabilities: {} });
await client.connect(new SSEClientTransport(new URL("https://nodusai-mcp-production.up.railway.app/sse")));

// Step 1 — get a session token at https://nodusai.app ($1 USDC)

// Step 2 — query the Oracle
const result = await client.callTool({
  name: "nodus_get_signal",
  arguments: {
    marketUrl:      "https://polymarket.com/event/...",
    sessionToken:   "your-session-token-from-nodusai.app",
    desiredOutcome: "YES", // optional
  }
});

Custom Python agent

from mcp.client.sse import sse_client
from mcp import ClientSession

async with sse_client("https://nodusai-mcp-production.up.railway.app/sse") as (read, write):
    async with ClientSession(read, write) as session:
        await session.initialize()

        # Get a session token at https://nodusai.app first ($1 USDC)
        result = await session.call_tool("nodus_get_signal", {
            "marketUrl":      "https://kalshi.com/markets/...",
            "sessionToken":   "your-session-token-from-nodusai.app",
            "desiredOutcome": "YES",  # optional
        })

Local development

git clone https://github.com/NodusAI-Your-Prediction-Broker/nodusai-mcp
cd nodusai-mcp
npm install

# Dev mode (mock oracle — no real API calls needed)
npm run dev:http

Test with:

curl http://localhost:3000/health
curl http://localhost:3000/info