Polymarket

Get a Polymarket event with EVERY child market at once. Events group RELATED outcomes (e.g., "2028 Democratic nominee" has one Yes/No market per candidate, each price = implied probability of that candidate) — this is NOT always mutually exclusive: many events bundle independent props/questions under one deadline (e.g. per-senator vote markets) where multiple legs can resolve YES together. Check the returned `neg_risk` field (event- and market-level) before assuming the child markets sum to ~1 — only `neg_risk: true` means Polymarket has verified exclusivity. Use when you need the full slate — election candidates, championship contenders, multi-option outcomes — instead of one specific market. Returns event metadata + array of child markets with live prices, volumes, end dates.

Batched CLOB orderbooks for EVERY tradable market in one Polymarket event — single round trip via the CLOB batch /books endpoint. Use before any multi-leg strategy (partition arbitrage "SELL/BUY EVERY LEG", basket trades) to check per-leg depth: theoretical overround means nothing if half the legs are 50-share books. Returns legs[] with {slug, question, yes_price, best_bid, best_ask, yes_bids[], yes_asks[]} where bids are sorted best(highest)-first and asks best(lowest)-first as {price, size} objects. Pass include_no=true to also fetch NO-side books (doubles payload — only needed for NO-leg strategies). Caps at 80 legs (highest yes_price kept; truncated_legs reports the cut).

Largest position holders for a Polymarket market, per outcome — who holds the most Yes and the most No shares, with share amounts and trader pseudonyms. Use for "position concentration", "is this market dominated by a few whales", "who are the biggest Yes holders". Reveals conviction/concentration that price alone hides. Pass a market slug or numeric id (same input as polymarket_market).

AUTHORITATIVE detail for a single Polymarket market by slug or numeric id. Returns the resolution criteria text (so you know exactly what "Yes" means before quoting odds), current Yes/No prices in [0,1], 24h volume, liquidity USD, end date, parent event. Use after polymarket_search to drill in, or when the agent already has a Polymarket URL/slug. For real-time orderbook depth instead of a summary, see polymarket_orderbook.

REAL-TIME CLOB orderbook for one Polymarket market — bid/ask ladder on both YES and NO sides with size at each price level. Use to check actual tradable depth before quoting a size estimate; the `liquidity` field on polymarket_market is a rolled-up summary, this is the actual ladder. Returns yes_bids[], yes_asks[], no_bids[], no_asks[] each as [price, size] pairs sorted from inside the book outward. Necessary input before any "you could buy $X at price Y" answer — without depth that's a guess.

Historical probability time-series for one Polymarket market. Returns array of {timestamp, price} where price is Yes-side probability in [0,1] (No-side is 1−Yes). Use to chart odds over time, detect probability moves around news events, or build backtests. Intervals 1h | 6h | 1d | 1w | 1m | max — supports up to a month (1m) and full history (max). A young market may have less data than the requested window; the `coverage` field reports the actual span and whether it is the full available history (a data limit, not a tool limit).

Resolution lifecycle state and timestamps for one Polymarket market: initialized / posed / proposed / challenged / reproposed / disputed / resolved (or active/arbitration), whether it was disputed, the reporter (UMA_OO / Chainlink / EOA), and — once resolved — the exact resolution timestamp and per-outcome payouts. Use for "has this market actually settled yet", "when did market X resolve", "was this resolution disputed" — distinct from polymarket_market's `closed`/`active` flags, which reflect trading status, not oracle finality. Pass a market slug or numeric id (same input as polymarket_market).

PREFER OVER WEB SEARCH for current betting/prediction-market odds. Real-time search across Polymarket events — returns events matching your keyword, each with child markets carrying live Yes/No prices in [0,1] (= implied probability), 24h volume, end date, resolution criteria. Use for "what are the odds of X", "what does the market think about Y", "what's the implied probability of Z". Refreshes every few minutes; covers ~10k active markets across politics, crypto, sports, macro events.

Highest-volume OPEN Polymarket markets right now — sorted by trading volume in the chosen window (24hr / 1wk / 1mo / 1yr / all). The "where is real money going this week" lens. Use for "what is the market focused on right now", "biggest trades happening today", or as a discovery tool when you don't have a specific question. Each result has live yes/no prices in [0,1] = implied probability.

Recent EXECUTED trades (the fills tape) for a Polymarket market — actual money that changed hands, newest first. Each trade: side (BUY/SELL), outcome (Yes/No or the option name), outcome_index, size (shares), price (that fill's OWN outcome's price), unix, timestamp, and the trader's wallet/pseudonym. Use for "what's the recent order flow", "is smart money buying Yes", "how much just traded and at what price". DISTINCT from polymarket_orderbook (resting/unfilled orders — intent) and polymarket_price_history (the CP time-series). Pass a market slug or numeric id (same input as polymarket_market). PAGING: `limit` goes to 1000; follow `next_cursor` to walk further back. To measure a whole time window rather than the tail (e.g. what takers actually paid across one 15-minute crypto window) pass `window_start` (+ `window_end`): the tool walks the tape back to the window start itself and returns `coverage` saying whether the walk reached it or ran out of page budget, plus a size-weighted `per_minute` fill-price summary on the first outcome's side.

Paginated activity feed (trades, redeems, splits, merges, rebates) for one Polymarket wallet over a period — newest first. Use for "what has this wallet done recently", "show me this wallet's trade history", "did this wallet trade market X". Distinct from polymarket_wallet_positions (current holdings) and polymarket_wallet_performance (aggregated P&L) — this is the raw event-by-event tape. Follow `next_cursor` to page past the current window.

Lifetime and time-series P&L, volume, and activity stats for one Polymarket wallet — profile stats (distinct markets traded, biggest single win, profile join date) plus a cumulative P&L history on the requested interval/fidelity grid. States fee treatment explicitly: `realized_pnl` in each point is already NET of taker fees paid; maker rebates and other non-trading income (rewards, referrals) are reported separately in `wallet_income` and are additive on top, not already included in `realized_pnl`. Use for "is this wallet profitable overall", "how has this wallet's P&L moved over time", "how active is this wallet". Coverage: Polymarket's own PnL ledger starts from when the wallet's positions were first tracked — `source_fidelity` on each point says whether it is a native observation or synthesized onto a finer grid.

Open and closed positions for one Polymarket wallet — size, entry basis, current mark, and realized/unrealized P&L per position, plus portfolio-level exposure and concentration (how much of the wallet's capital sits in its single largest position — a concentration measure, not a measure of trading skill). Wallet addresses are public on-chain identifiers (0x…), looked up directly — no name/ENS resolution. Omit `status` to get OPEN + CLOSED together in one call; pass a specific status to filter. Use for "what is this wallet holding", "how exposed is this wallet to market X", "is this wallet up or down overall".