Service · Cross-platform

Polymarket and Kalshi disagree on the same event. Capture the gap.

The same election, the same Fed decision, the same BTC threshold - priced differently on two venues. Cross-platform arbitrage bots match the events, size against both books, and fire matched legs across both APIs. Built for capital that can operate in both jurisdictions (Kalshi US-legal, Polymarket internationally accessible).

210 /mo
"polymarket kalshi arbitrage" US search volume
src: DataForSEO May 2026
$129 CPC
"kalshi vs polymarket arbitrage" CPC
src: DataForSEO May 2026
$183 CPC
"polymarket kalshi arbitrage finder" - highest CPC in cluster
src: DataForSEO May 2026
2-8¢
Typical price gap on shared political/macro events
src: 90-day snapshot study
How it works

Three problems to solve, in order.

Cross-venue arb sounds simple. Implementation is hard because the two venues speak different languages.

01 · MATCH

Identify shared events across venues. "Will Trump win in 2028?" on Polymarket maps to PRES-2028-DEM series on Kalshi. We use semantic matching (LLM-assisted) + human review for ambiguous cases.

02 · NORMALIZE

Different costs, different sizes. Kalshi has $0.07 minimum fees per trade. Polymarket has gas (~$0.005). We normalize to "net price after costs" before declaring an edge.

03 · EXECUTE

Two APIs, one execution path. Polymarket CLOB + Kalshi Trade API in a unified router. Legs fire simultaneously with partial-fill unwind logic.

Config

Cross-venue config sample.

Geographic enforcement is in code. The bot won't fire Polymarket legs if the user is in a blocked jurisdiction.

cross-arb.yaml config
# Cross-venue arbitrage - production config
strategy: "cross-venue-arb"

venues:
  - id: "polymarket"
    enabled_for_geos: ["IN", "IL", "AE", "HK", "CA-non-ON"]
  - id: "kalshi"
    enabled_for_geos: ["US"]

event_matching:
  matcher: "semantic-llm"
  similarity_threshold: 0.85
  manual_review_below: 0.85
  human_approved_pairs_only: false  # strict mode toggle

trigger:
  min_edge_cents_net: 2.5      # after fees + gas
  min_orderbook_depth_usdc: 1_500   # both venues combined

execution:
  legs_in_parallel: true
  slippage_bps_per_leg: 50
  partial_fill_policy: "unwind"
  geographic_check: true            # enforced before each fire

risk:
  max_position_per_event_usdc: 8_000
  daily_drawdown_stop_usdc: 2_500
Honest framing

Why this is harder than it sounds.

i
Event matching errors are expensive.

If you accidentally arb "Will Trump win the primary?" against "Will Trump win the general?", you've taken a directional bet without knowing it. Semantic matching needs a human review layer until you've validated the pair history. We bake that in.

i
Resolution dates rarely match perfectly.

Kalshi might settle on a date Polymarket interprets differently. We track resolution conditions verbatim per market and refuse to pair markets where the criteria differ materially.

!
Two venues = two custodies = two operational risks.

Funds in Kalshi USD account. Funds in Polymarket USDC on Polygon. Rebalancing between them takes hours and costs fees. We surface treasury imbalance in the dashboard.

×
No geographic workarounds.

If you're US-based, you trade Kalshi only. If you're in a Polymarket-accessible jurisdiction, you can trade both. We won't ship cross-venue bots to clients whose geographic eligibility we can't verify.

Where this works

Categories with persistent cross-venue gaps.

Politics 2-8¢

Presidential & state elections

Polymarket leans crypto-native sentiment, Kalshi leans US-resident sentiment. Persistent 2-4¢ gaps on major races, 5-8¢ on long-tail.

Macro (Fed, CPI) 0.5-2¢

Fed decisions, inflation prints

Tighter gaps but very frequent. Kalshi's Fed series have weekly resolutions - high turnover. Where most cross-venue volume happens.

Crypto thresholds 1-5¢

BTC / ETH price levels

Polymarket has deeper liquidity (crypto-native users). Kalshi's prices often lag during volatile sessions. Cross-venue gaps widen.

Budget bracket

Where cross-venue arb lands.

Single-category arb
$25k-$50k · 6-8 weeks

Polymarket↔Kalshi on one category (e.g., politics) with manual event-pair approval.

  • Both venues, one category
  • Manual pair approval workflow
  • Geographic enforcement
  • Grafana dashboard
  • 60-day warranty
Get started

Want to arb both venues?

Send us your jurisdiction(s) of operation and target categories. Cross-venue projects start with a compliance + jurisdiction review - typically a 45-minute call before we quote.