Skip to content

Kalshi and Polymarket,
matched into one real-time API

Two venues, one call

dino.markets matches equivalent markets across Kalshi and Polymarket and streams the best available line over one WebSocket, sized to fillable depth, with arbitrage signals computed net of fees. Settlement rules are checked before anything reaches you. Callable via MCP.

France / BrazilSpread:1.84%

Built to build on

One WebSocket for the live feed and REST for history, with an MCP server so any agent can call it. Subscribe to a category or the full feed and act the moment a line opens. Every field is documented and typed, so you skip the polling loop and the matching code.

// one connection, both venues
centrifuge = Centrifuge("wss://stream.dino.markets/connection/websocket")
centrifuge.connect()
▸ Spurs / Knicks +1.84% $2,600

The honest best line

Every match arrives already priced. You see the spread after fees and the size the books can fill, with a confidence flag you can filter on. Low-confidence matches never reach your feed, so what you act on is what we already trust.

Spurs vs Knicks
1.84%
Kalshi logoSpurs · yes$0.54depth $3,100
Polymarket logoKnicks$0.44depth $2,900
Executable size$2,600

Matched markets

FED·Wed, Sep 16
Order book →
KalshiPolymarketGap
No change
65¢ask63¢ask2¢
Hike 25bps
35¢ask36¢ask1¢
Cut 25bps
askask
K 4.3M · PM $19.1M
MMA·Sat 5:00 PM ET
Order book →
KalshiPolymarketGap
ian machado garry logoIan Machado Garry
26¢ask25¢ask
1¢
islam makhachev logoIslam Makhachev
76¢ask76¢ask
level
K 666.5K
FED·Wed, Oct 28
Order book →
KalshiPolymarketGap
No change
70¢ask68¢ask2¢
Hike 25bps
23¢ask25¢ask2¢
Cut 25bps
askask1¢
K 116.6K · PM $541.8K
NFL logoAm. Football·Wed 8:20 PM ET
Order book →
KalshiPolymarketGap
new england patriots logoNew England Patriots
37¢ask34¢ask
3¢
seattle seahawks logoSeattle Seahawks
66¢ask67¢ask
1¢
K 126.7K · PM $127

Who dino.markets is for

01

Trading bots

Subscribe to a category or the full feed, act on the open market, and size to max_stake. You skip the polling loop and the matching code.

02

Quant and arb desks

Signal you can filter by confidence and settlement flags, already net of fees, ready to size on your own capital.

For coding agents too

Paste this into Claude or Cursor to read the docs and pull live markets with your key.

Read https://dino.markets/llms.txt then use my dino.markets API key sk_live_YOUR_KEY to list today's live cross-venue markets and show any confirmed arbitrage.

The MCP server is on every plan, including Free. See the docs

Pricing

Free
$0
  • Matched feed across both venues
  • REST API, about two minutes behind
  • 10,000 requests a month at 10 req/s
  • 14-day price history
  • Free MCP access
  • 1 real-time connection on the curated sample channel (the full market stream starts at Basic)
Real-time
Basic
$30 / mo
  • Everything in Free
  • Real-time WebSocket push
  • Subscribe by category or to the full feed
  • 3 concurrent connections
  • 1,000,000 requests a month at 10 req/s
Premium
$100 / mo
  • Everything in Basic
  • 10 concurrent connections
  • 5,000,000 requests a month at 25 req/s
Pro
$200 / mo
  • Everything in Premium
  • Raw quote-level data
  • Early candidate signals
  • Priority support

FAQ

Is this financial advice?

+
No. We report the numbers, like “these two prices form a 1.84% gap at $2,600 of fillable size, net of fees.” We never say who will win. You hold the accounts and the capital on both venues and trade yourself.

Are these a sure thing?

+
No. A gap can close before you act, and a settlement rule can differ between venues. So every signal shows the fillable size and the fee-adjusted spread, plus a resolution-confidence flag you can filter on. Low-confidence matches never reach your feed.

Do you place trades for me?

+
No. We are a data API, not a broker. You trade on Kalshi in USD and Polymarket in USDC on Polygon yourself.

How is this different from Prediction Hunt or Oddpool?

+
They cover more venues and a broader catalog. We focus on matching Kalshi and Polymarket across sports, weather, crypto, and economics, with explicit entity and settlement-rule confidence on every pair.