Quickstart

Get from zero to your first signal in three steps.

Get an API key

Request access to get an API key, or use the Flex tier to pay per request via x402 with no account. See the plans page for all tiers.

Make your first call

Fetch the top trending markets in a single request:

curl -H "x-api-key: YOUR_KEY" \
  "https://api.scanna.xyz/hot?limit=5"

Read the response

Each market comes back with a composite heat_score, whale activity, and human-readable signals:

{
  "markets": [
    {
      "market_id": "0xfb1835...",
      "question": "Will Manchester City win the 2025-26 EPL?",
      "heat_score": 0.6,
      "volume_spike_x": 1.2,
      "hours_to_resolve": 1566,
      "liquidity_usd": 388845,
      "whale_activity": {
        "count": 3,
        "total_notional": 4678,
        "direction": "buy"
      },
      "volume_24h": 19274,
      "price_yes": 0.09,
      "flow_bias": "buy",
      "tags": [],
      "end_date": "2026-05-27T00:00:00Z",
      "signals": [
        "3 whale buys totaling $4,678",
        "Net buyers dominating",
        "Trade frequency above baseline"
      ]
    }
  ],
  "updated_at": "2026-03-22T18:06:24.950Z"
}

Go deeper

Drill into a single market with POST /heat, score a market or wallet with Analytics, or pull an ML prediction. When you're ready to wire it into a bot, copy a working integration example.

Tip

A good first lap: GET /health (no auth) → GET /hot → POST /heat on an interesting market → GET /predict for a model probability.

Suggested testing order

When evaluating the full surface, work through it roughly in this order:

Health — GET /health confirms the service is up (no key needed).
Trending — GET /hot for the ranked feed.
Deep dive — POST /heat for one market's full signal breakdown.
Prediction — GET /predict (resolved) or GET /live-predict (active).
Risk & signals — GET /risk-score, GET /signals.
Wallets — GET /smart-money, GET /wallet/{address}.
Cross-venue — GET /divergence.
Export — GET /export for training data.
Monitoring — GET /metrics for cache, DB, and model status.