> ## Documentation Index
> Fetch the complete documentation index at: https://docs.agentkey.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Prediction market data for AI agents

> Polymarket and Kalshi events, markets, prices, order books, trades, volume, open interest, smart money and cross-platform matching through AgentKey.

Prediction markets put a live probability on almost anything, which makes them
a rich signal for an agent — if it can read them. Polymarket and Kalshi expose
different APIs, use different identifiers and quote in different units, and
matching the same question across the two is manual work. Through AgentKey your
agent reads both platforms the same way — `find_tools`, `describe_tool`,
`execute_tool` — with Surf behind the scenes, including analytics that already
line the two up.

## What you can ask

* *"What are the odds on Polymarket for this event, and how have they moved this month?"*
* *"Find the same election market on Polymarket and Kalshi and tell me the price spread."*
* *"Which prediction markets have the most volume this week?"*
* *"What is smart money betting on in the politics category?"*
* *"Show the order book for this Kalshi market."*
* *"Who are the top Polymarket traders by realized PnL?"*
* *"这个钱包在 Polymarket 上持有哪些仓位？"*
* *"Which crypto-category markets move together?"*

## What's covered

**Discovery**

* Browse and search markets across platforms by platform, category, status,
  keyword and smart-money direction, with volume, open interest, trade counts
  and smart-money signals per result
* Polymarket events with nested markets and two-sided outcomes; Polymarket
  markets by slug
* Kalshi events with nested markets; Kalshi markets with price, volume and
  status

**Prices and order books**

* Polymarket price history as one point per interval, or as OHLCV candles for
  charting
* Kalshi price history as daily or hourly OHLC, or the latest price from
  recent trades
* Historical order book snapshots — raw bid/ask depth on Polymarket, yes-side
  bid/ask levels on Kalshi

**Volume, open interest and trades**

* Volume history on both platforms, and a YES/NO volume split per hour, day or
  week on Polymarket
* Daily open interest history on both platforms
* Individual trades — on Polymarket filtered by market, wallet, outcome,
  minimum size and time range, including splits, merges and redemptions; on
  Kalshi filtered by taker side, minimum contracts and date range

**Traders and smart money**

* Polymarket leaderboard ranked by realized PnL, volume or trade count, with
  wallet addresses
* Smart-money positioning per market — net bullish or bearish signal and
  conviction — and the individual large smart-wallet trades behind it
* Open positions of one wallet on Polymarket

**Analytics and cross-platform**

* Platform-wide totals of volume, open interest and market count; trends, top
  markets, momentum distribution and per-market momentum signals, by category
  or across everything
* 30-day price correlations between Polymarket markets within a category,
  covering the highest-volume markets
* Matched Polymarket–Kalshi pairs for the same event with live price spread,
  and a daily volume and open interest comparison for a matched pair

## How to use it

Pass the request in plain language and let the router pick the tool:

```text theme={null}
find_tools(q="Polymarket odds and price history for the US election market")
find_tools(q="Kalshi 和 Polymarket 上同一事件的价差")
find_tools(prefix="crypto/prediction")    # browse everything for prediction markets
describe_tool("<returned Provider/Operation>")
execute_tool("<Provider/Operation>", { ...params })
```

Both English and Chinese phrasing work. Always `describe_tool` before
executing — Polymarket tools key on condition ids and slugs, Kalshi tools on
event and market tickers, and the per-call cost is shown there.

> [!NOTE]
>
> **Building with code?** Prediction market data is available through the
> catalog today; there is no dedicated Scenario Method yet. See
> [Scenario Methods](/api-reference/scenario/overview) for the fixed-contract
> API and the methods that exist.

## FAQ

> **Which platforms are covered?**
>
> Polymarket and Kalshi, individually and through the cross-platform search,
> analytics and pair-matching tools that span both.

> **Can the agent place bets?**
>
> No. Every tool is read-only: markets, prices, order books, trades and
> positions are observed, never created.

> **Is historical data available?**
>
> Yes — price history and candles, order book snapshots, volume and open
> interest series and trade records are all time series. Market listings,
> leaderboards and smart-money positioning are fetched at request time.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.