> ## Documentation Index > Fetch the complete documentation index at: https://docs.graphadvocate.com/llms.txt > Use this file to discover all available pages before exploring further. # Kalshi derived signals > Three paid endpoints — consensus trend, cross-source spread, and live-sports edge. JOINs and derived scores passthrough APIs can't replicate, on top of Kalshi's fully public REST. Three x402 endpoints for **US-regulated prediction-market intelligence**. Kalshi's REST API is fully public — but the value here is in the *derivation*: slopes, cross-source JOINs, and momentum-vs-market signals an agent would otherwise have to compute itself. ## What you get | Endpoint | Cost | Returns | | -------------------------------- | -----: | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `POST /kalshi/consensus-trend` | \$0.05 | Slope-per-hour over 24h + 3d, acceleration signal, volatility band, days-to-resolve — derived from Kalshi's *unique* `forecast_history` | | `POST /kalshi-polymarket/spread` | \$0.05 | Cross-source spread between Kalshi and Polymarket on the same topic, with arbitrage direction. **JOIN that single-source passthrough APIs structurally can't return** | | `POST /kalshi/sports-live-edge` | \$0.05 | Play-by-play momentum vs market candle reaction → latency-arb signal for live Football / Basketball / Soccer / Hockey / Baseball / WNBA markets | Per-call payment in USDC on Base via x402. No subscription, no API key, no free tier. ## Why derived over raw Kalshi exposes \~50 public no-auth endpoints across 10,821 series (Entertainment 2,445 · Sports 2,172 · Politics 1,986 · Elections 1,404 · Economics 563 · Financials 516 · …). An agent can read them directly — but every consuming agent then re-implements the same math: * Fit a slope through `forecast_history` points to detect regime change * Match Kalshi events against Polymarket equivalents by keyword + price proximity * Compute momentum from play-by-play events and compare to market reaction lag Graph Advocate does it once, exposes the answer. The cross-source `/spread` endpoint in particular is the **defensible play**: passthrough APIs return one source at a time and can't JOIN — so the cross-venue arbitrage signal stays valuable even as more upstream sources come online. ## Sample: consensus trend for an event ```bash theme={null} curl -X POST 'https://graphadvocate.com/kalshi/consensus-trend' \ -H 'Content-Type: application/json' \ -d '{"event": "KXELONMARS-99"}' ``` Response (truncated): ```json theme={null} { "kalshi_event_ticker": "KXELONMARS-99", "event_title": "Will Elon Musk visit Mars in his lifetime?", "category": "World", "consensus_probability_now": 0.18, "slope_per_hour_24h": -0.0004, "slope_per_hour_3d": -0.0001, "acceleration_signal": -0.0003, "interpretation": "accelerating-down", "volatility_24h_stdev": 0.012, "days_to_resolve": 3287.4, "markets_in_event": 1, "history_points_analyzed": 142, "kalshi_source": "https://api.elections.kalshi.com/trade-api/v2/events/KXELONMARS-99/forecast_history", "agent_note": "Forecast history is Kalshi's published consensus probability over time. Use slope+acceleration to detect regime changes before they're priced in. Pair with /kalshi-polymarket/spread for cross-source arbitrage." } ``` **How to use it.** A positive `acceleration_signal` means the slope is *steepening upward* — the market is converging on YES faster than it was 3 days ago. Negative means the opposite. `interpretation` collapses the signal into one of `accelerating-up`, `accelerating-down`, `stable`, or `insufficient-history` so an agent can branch without parsing. ## Sample: cross-source spread ```bash theme={null} curl -X POST 'https://graphadvocate.com/kalshi-polymarket/spread' \ -H 'Content-Type: application/json' \ -d '{"topic": "fed rate", "limit": 5}' ``` Response: ```json theme={null} { "topic_keyword": "fed rate", "kalshi_candidates": 3, "polymarket_candidates": 2, "pairs": [ { "kalshi_ticker": "KXFED-25DEC-CUT25", "kalshi_yes_mid": 0.62, "polymarket_market_slug": "fed-rate-cut-december", "polymarket_yes_mid": 0.58, "spread_yes_kalshi_minus_poly": 0.04, "spread_bps": 400, "arbitrage_direction": "long-poly-short-kalshi" } ], "agent_note": "Spread > 200bps in either direction is a candidate arbitrage; verify the two markets actually resolve on the same condition before sizing. Cross-source data single-source APIs can't return in one call.", "sources": { "kalshi": "https://api.elections.kalshi.com/trade-api/v2", "polymarket": "token-api proxy" } } ``` **Important:** the pairing uses keyword overlap + price-proximity scoring. **You must verify the two markets actually resolve on the same on-chain condition before sizing** — otherwise the "arbitrage" is just basis risk. ## Sample: live sports edge ```bash theme={null} curl -X POST 'https://graphadvocate.com/kalshi/sports-live-edge' \ -H 'Content-Type: application/json' \ -d '{ "milestone": "", "market": "" }' ``` The `milestone` is Kalshi's identifier for a specific live game; `market` is an optional ticker for the market whose candlesticks you want compared against the play-by-play. Response includes `momentum_score_last_5_events`, `market_reaction_pct_last_hour`, and a `latency_arbitrage_signal` of one of `upside-lag-likely`, `downside-lag-likely`, `market-tracking-stats`, or `null` (insufficient data). **How to use it.** When `momentum_score` is strongly one-way (≥0.6 or ≤0.2) but `market_reaction_pct` is near zero, the market hasn't fully priced the play-by-play yet — latency-arb window. Verify orderbook depth before sizing — low liquidity will eat the edge. ## How it differs from `/polymarket/*` and `/hyperliquid/*` These three Kalshi endpoints are **derived signals** on top of a *fully public* upstream — different model than the polymarket/hyperliquid trader-intelligence endpoints which proxy through gated data sources. Kalshi's API is free for anyone to read; Graph Advocate's value here is in the math: * `/polymarket/risk` — wallet-type detection (binary signal from on-chain probe) * `/hyperliquid/risk` — liquidation rate + funding burn (derived from per-trade ledger) * `/kalshi/consensus-trend` — slope of Kalshi's *unique* forecast percentile history * `/kalshi-polymarket/spread` — JOIN that no single-source API can return * `/kalshi/sports-live-edge` — momentum-vs-market reaction lag If you want to inspect the routing for free, send a plain-English question to `POST /` (A2A JSON-RPC) and Graph Advocate will name the right `kalshi/*` endpoint and hand back the exact paid curl — without charging. ## Source [`kalshi.py`](https://github.com/PaulieB14/graph-advocate/blob/main/kalshi.py) — the three derivation functions and how they fetch from Kalshi's public REST and Polymarket data sources.