Polymarket Node

The Polymarket node works with prediction markets on Polymarket ↗ — the world's largest prediction market platform. It has six modes across two families: discovery (browse / search / lookup) finds markets and their identifiers via the Gamma API, and market data (snapshot / history / depth) fetches live pricing, price history, and orderbook depth for a single market via the CLOB API. Both APIs are free and require no API key. The output shape is discriminated on mode.

Discovery Modes (Gamma API)

These modes return a markets[] array. Use them to find markets and their identifiers (conditionId, tokenIds, slug).

Browse (default)

Filter-based discovery using the Gamma /events/keyset endpoint. All fields are optional — with no filters it returns the top markets sorted by 24h volume.

Search

Free-text search using /public-search. Matches against market question text.

Lookup

Direct fetch by event slug or numeric event ID.

Note: Lookup expects an event slug or event ID. Individual market slugs from discovery output (markets[i].slug) do not round-trip here — Gamma returns 0. To drill into a discovered market, chain into a market-data mode (snapshot/history/depth) via its conditionId/tokenId, or use the numeric eventId for lookup.

Multi-outcome events (e.g. "How many rate cuts in 2026?") return multiple rows in markets[] — one per possible outcome.

Market-Data Modes (CLOB API)

These modes fetch detailed data for a single market per execution via the CLOB API. They share one Market field that accepts four formats, auto-detected at runtime: a condition ID (0x… + 64 hex), a CLOB token ID (~60–90 digit decimal), a slug (fed-rate-cut-2026), or a Polymarket URL. Flip the field to f(x) mode to pass an identifier from an upstream node — e.g. {discovery.markets[0].conditionId} or {loop.item.tokenIds.yes} — which is how you fan a discovery result (or a loop) into per-market data.

snapshot accepts a condition ID and returns both YES + NO. history and depth operate on a single outcome token — pick YES/NO in the UI, or pass a token ID directly.

Snapshot

Current pricing and orderbook summary for a market.

Output: snapshot.market (metadata), snapshot.yes / snapshot.no (each { tokenId, price, bid, ask, mid, spread }, or null if the input resolved to a single opposite token).

History

Price timeseries for a single outcome token, by relative lookback.

Output: history.points[] — [{ timestamp, price }] (aligned to the UTC interval boundary) — plus history.tokenId and history.fidelityMinutes.

Note: the most recent point is the in-progress (not-yet-closed) candle — its timestamp is the request time, not a boundary. (Trimming it to the last closed candle is a planned follow-up.)

Depth

Full orderbook for a single outcome token.

Output: depth.bids[] / depth.asks[] ([{ price, size }], sorted), depth.lastTradePrice, depth.tickSize, depth.minOrderSize, depth.totalBidSize, depth.totalAskSize.

Workflow Examples

Monitor Top Political Markets

Find the most active political prediction markets each morning and summarize them with an LLM.

Configure the Discovery node with tag=politics, sortBy=volume_24hr, limit=10. The LLM receives the markets array and formats a digest of the most liquid political bets with their current implied probabilities.

Search and Alert on New Markets

Search for markets matching a topic and alert when new ones appear above a liquidity threshold.

Use search mode with query="fed rate cut 2026". The Function node filters for markets with liquidity > 10000. The Conditional node checks if filtered.markets.length > 0 before sending a Telegram alert.

Lookup a Specific Market

Fetch all outcome markets for a known event by slug and pass identifiers to a downstream node.

Pricing & Credits

Each Polymarket call costs a flat 2 credits, regardless of mode (discovery or market data) or number of markets returned.

Output

The output is discriminated on mode. Discovery modes return markets[]; market-data modes return a snapshot / history / depth object instead. For discovery, markets are flattened from events — each child market (each possible outcome) is a separate row in markets[].

Market Object Structure

Each item in markets[] has:

Example Output (Browse Mode)

{
  "mode": "browse",
  "markets": [
    {
      "question": "Will the Fed cut rates in September 2026?",
      "slug": "will-the-fed-cut-rates-september-2026",
      "eventId": 98712,
      "conditionId": "0xabc123...",
      "tokenIds": { "yes": "71321...", "no": "71322..." },
      "outcomePrices": { "yes": 0.72, "no": 0.28 },
      "volume": 4500000,
      "volume24hr": 320000,
      "liquidity": 180000,
      "endDate": "2026-09-30T00:00:00Z",
      "tags": ["economics", "fed", "rates"]
    }
  ],
  "totalMatched": 1,
  "metadata": {
    "source": "polymarket-gamma",
    "endpoints": ["/events/keyset"],
    "requestedAt": "2026-05-07T09:00:00Z",
    "dataPoints": 1,
    "creditsUsed": 2
  }
}

Common Errors

Next Steps

• LLM Node — Feed market data and implied probabilities into an LLM to generate trading analysis.
• Function Node — Filter, sort, or transform the markets[] array with custom JavaScript.
• Conditional Node — Branch your workflow based on odds thresholds or liquidity levels.