Polymarket & Kalshi Prediction Market Scanner – Cross-Platform Arbitrage Gaps

Scrape active Polymarket and Kalshi prediction markets into one clean, unified dataset and automatically match the same event across both platforms to surface YES-price gaps – a starting point for prediction market arbitrage, odds comparison, research dashboards and alerts.

• ✅ Polymarket (Gamma API) + Kalshi (official Trade API v2) – public market data only, no login, no API keys
• ✅ One schema for both platforms: price, 24h volume, liquidity, end date, URL
• ✅ Cross-platform matching (keywords + end-date rules + fuzzy similarity) with match_confidence
• ✅ price_gap = Polymarket YES − Kalshi YES for every matched pair
• ✅ Filters: category, minimum 24h volume, max markets, matched-pairs-only mode
• ✅ Pay-per-event pricing, respects your maximum spend limit

What can I use it for?

• Find Polymarket vs Kalshi price differences on the same question (elections, sports, esports, awards, politics…)
• Build a prediction-market odds comparison table or Google Sheet
• Monitor the most active prediction markets by 24h volume
• Feed research / quant models with normalized market probabilities

⚠️ A price gap is not guaranteed arbitrage. Resolution rules, end dates, fees, spreads, liquidity and platform/jurisdiction access differ. Always read both markets' rules before trading. This Actor is a data tool, not financial advice.

Input

Field	Type	Default	Description
categories	string[]	[]	Keep markets whose category/tags contain any of these words (e.g. Politics, Sports, Crypto, Elections). Empty = all.
minVolume24h	integer	1000	Minimum 24h volume (Polymarket: USD; Kalshi: contracts, $1 notional each).
maxItemsPerPlatform	integer	500	Max markets kept per platform (highest 24h volume first).
onlyMatchedPairs	boolean	false	Output only matched pairs (one row per pair, sorted by largest gap).
matchThreshold	string (0–1)	0.6	Minimum match confidence.
maxEndDateDiffDays	integer	7	Max end-date difference allowed for a match.

Output fields

platform, market_id, question, category, yes_price, no_price, volume_24h, liquidity, end_date, url, fetched_at, matched_market_id, matched_platform, price_gap, match_confidence (+ matched_question, matched_url on matched rows).

Notes:

• Polymarket yes_price = first outcome price from Gamma outcomePrices (for head-to-head sports markets the first outcome is the first-named team).
• Kalshi yes_price = mid of best YES bid/ask (falls back to last trade); no_price = 1 − yes.
• Kalshi has no liquidity field; liquidity holds open interest (contracts) for Kalshi rows.

Example output (real row from a test run on 2026-10-16)

{
  "platform": "polymarket",
  "market_id": "559653",
  "question": "Will Alexandria Ocasio-Cortez win the 2028 Democratic presidential nomination?",
  "category": "United States",
  "yes_price": 0.166,
  "no_price": 0.834,
  "volume_24h": 10385.92,
  "liquidity": 367883.6745,
  "end_date": "2028-11-08T04:59:00Z",
  "url": "https://polymarket.com/event/democratic-presidential-nominee-2028",
  "fetched_at": "2026-10-16T03:20:49Z",
  "matched_market_id": "KXPRESNOMD-28-AOC",
  "matched_platform": "kalshi",
  "price_gap": 0.031,
  "match_confidence": 0.807,
  "matched_question": "Will Alexandria Ocasio-Cortez be the Democratic Presidential nominee in 2028?",
  "matched_url": "https://kalshi.com/markets/kxpresnomd"
}

More real examples: sample_output_matched.json.

Pricing (pay per event)

Event	When	Suggested price
market_row	a normalized market row without a match	$0.001
matched_gap	a row/pair matched across Polymarket and Kalshi	$0.005

The Actor stops pushing results as soon as your maximum total charge would be exceeded.

How matching works

1. Normalize text (accents, “U.S.” → “us”), drop stop-words, index Kalshi markets by keywords.
2. Candidates must share ≥2 distinctive keywords and have end dates within maxEndDateDiffDays.
3. Score = 0.6·keyword Jaccard + 0.25·rapidfuzz token-set ratio + 0.15·date proximity; penalized when numbers/strikes differ or when market types conflict (spread vs moneyline, nomination vs election…).
4. Greedy one-to-one assignment by best score.

Run locally

pip install -r requirements.txt
mkdir -p storage/key_value_stores/default
echo '{"onlyMatchedPairs": true, "maxItemsPerPlatform": 1500}' > storage/key_value_stores/default/INPUT.json
python -m src            # or: apify run
./test_ppe_local.sh      # simulate pay-per-event with a $0.05 spend limit

Data sources

• Polymarket Gamma API: https://docs.polymarket.com
• Kalshi Trade API v2: https://docs.kalshi.com

Not affiliated with Polymarket or Kalshi.