# SimpleFunctions — Full Reference (Live)

Generated: 2026-04-27T05:17:38.862Z
Every number in this document is pulled from live DB at request time. Cached 5 min at edge. No stale marketing copy.

---

## What SimpleFunctions is

Prediction market data and AI-powered analytics for agents. Kalshi + Polymarket compressed into structured real-time context. Instead of web search returning narrative paragraphs, agents receive probabilities: "Iran military action 18¢ (+5¢), GeoRisk 33/100, Oil +3.2%". One API call, ~800 tokens, refreshed every 15 min, free.

## Live state snapshot

- Markets tracked: 32,467
  - Kalshi: 22,620
  - Polymarket: 9,847
- Regime: Steady — no dominant macro signal
- SF Index v2:
  - Disagreement: 37/100
  - GeoRisk: 33/100
  - Breadth: +0.25
  - Activity: 100/100
- Cross-venue pairs: 1,396
  - Arb-ready (conf ≥0.75 AND |Δt| ≤90d): 645
- Trade ideas: 6 active

## Right now (10 salient items)

- [residual] 29.5σ volume @ 12¢ — federal funds rate be above 4.00%?
  Follow: https://simplefunctions.dev/api/agent/world?op=explain&item=s-om4kw5
- [residual] 12.5σ volume @ 45¢ — federal funds rate be above 3.50%?
  Follow: https://simplefunctions.dev/api/agent/world?op=explain&item=s-om3y0x
- [consensus_break] RTX 5090 compute per hour price be above?: 37 markets deeply split (σ=45¢, range 94¢)
  Follow: https://simplefunctions.dev/api/agent/world?op=dispersion
- [divergence] What will WTI Crude Oil (WTI) hit in … overshooting What will WTI Crude Oil (WTI) hit in … — 17¢ gap [Oil & energy]
  Follow: https://simplefunctions.dev/api/agent/world?op=explain&item=s-ev7h48
- [consensus_break] # of seats won by PH in Colombia Senate Elec…: 3 markets deeply split (σ=44¢, range 94¢)
  Follow: https://simplefunctions.dev/api/agent/world?op=dispersion
- [consensus_break] James Bialota be the Republican nominee for …: 3 markets deeply split (σ=43¢, range 92¢)
  Follow: https://simplefunctions.dev/api/agent/world?op=dispersion
- [divergence] How many ships will Iran successfully… lagging there be more than 60 transit calls t… — 15¢ gap [Iran / Hormuz crisis]
  Follow: https://simplefunctions.dev/api/agent/world?op=explain&item=s-ekn8os
- [resolving_soon] 95% yes in 13h — WTI front-month settle oil price be >89.99?
  Follow: https://simplefunctions.dev/api/public/market/KXWTI-26APR27-T89.99
- [resolving_soon] 95% no in 13h — WTI front-month settle oil price be >100.99?
  Follow: https://simplefunctions.dev/api/public/market/KXWTI-26APR27-T100.99
- [resolving_soon] 95% no in 13h — WTI front-month settle oil price be >101.99?
  Follow: https://simplefunctions.dev/api/public/market/KXWTI-26APR27-T101.99

## Top arb-ready cross-venue pairs

- Who will win Hart Memorial Trophy?: Alex Debrincat — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-ADEB
  Poly:   0xd9e01142203b0f1a829ac4b1e315d65ab7a237c2ab1b495ca5d881fd75b5544e
- Who will win Hart Memorial Trophy?: Adam Fox — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-AFOX
  Poly:   0x52092b7f17bb7e81ee6e8a096a3359474d4c51dc4324256056f916f7eca1031a
- Who will win Hart Memorial Trophy?: Adrian Kempe — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-AKEM
  Poly:   0x661ace963402a81529c195e3e4039e5040d240215184f8a998ad53a5a13cef9b
- Who will win Hart Memorial Trophy?: Auston Matthews — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-AMAT
  Poly:   0xebd6c8ce5c29f30a4837cdcd9a5f4cfa99501e4751e345fc4b244b2ad2cfe4af
- Who will win Hart Memorial Trophy?: Alexander Nikishin — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-ANIK
  Poly:   0x200622a87d7bef80fbad4d7e79918599e1f35b009c0a20ec3c2bcd2c27e1b18f
- Who will win Hart Memorial Trophy?: Artemi Panarin — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-APAN
  Poly:   0xaf76f1101ff8d67ede431bf47ba3797dc93534eedce9c7601025ae72734ff377
- Who will win Hart Memorial Trophy?: Andrei Vasilevskiy — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-AVAS
  Poly:   0x7cdc2fd58ba14920b36fd5368c6e688d01d2bdd0eafbde8c2677b4a6fd2bd388
- Who will win Hart Memorial Trophy?: Brady Tkachuk — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-BTKA
  Poly:   0x21e12abbe8356852466f27d26fec4c92094875a0ee81f2beb2b602caed41f4f2
- Who will win Hart Memorial Trophy?: Connor Bedard — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-CBED
  Poly:   0x8f22c12f69b15124af1af9a19ccc1bcf63edf32b587cd5b0d5684bfcab8d4a59
- Who will win Hart Memorial Trophy?: Connor Hellebuyck — conf 0.98, |Δt| 1d
  Kalshi: KXNHLHART-26-CHEL
  Poly:   0x9f2b96913520f451be509134fdf0a61a3f815fd11842483940f31c971186898d

## Endpoints

### /api/agent/world — Salience snapshot (SPEC-10)

Every call returns a Snapshot: { asOf, sinceBaseline, regime, salient[], childRegions[] }.
Salient items are type-tagged: regime_shift | divergence | resolving_soon | residual | novel | consensus_break.
Each item carries 2-3 pull ops — follow-up URLs the agent can GET to zoom.

Grammar:
  GET /api/agent/world                            → root snapshot
  GET /api/agent/world/iran                       → drill (path extends hierarchy)
  GET /api/agent/world/iran/hormuz                → drill deeper
  GET /api/agent/world/iran?op=history&dt=6h      → probability timeline
  GET /api/agent/world/iran?op=catalyst&window=48h → upcoming resolutions
  GET /api/agent/world/fed-rate?op=dispersion     → intra-topic consensus/disagreement
  GET /api/agent/world?op=trail&from=KXHORMUZ...  → linkage graph walk
  GET /api/agent/world?op=explain&item=s-abc123   → hypothesis candidates for a salient item

Modifiers (any op): since=12h, depth=0-3, limit=1-30, format=json|markdown.

### /api/agent/world/delta — Incremental (30-50 tokens)

  GET /api/agent/world/delta?since=1h
  GET /api/agent/world/delta?since=2026-04-17T00:00Z

### /api/public/cross-venue/pairs — Kalshi × Polymarket (SPEC-09)

Presets + explicit filters. Every consumer uses the same query language.

  GET /api/public/cross-venue/pairs?preset=arb      → conf ≥0.80, |Δt| ≤90d (645 rows)
  GET /api/public/cross-venue/pairs?preset=world    → conf ≥0.75, |Δt| ≤180d
  GET /api/public/cross-venue/pairs?preset=detail   → conf ≥0.60, no dt cap
  GET /api/public/cross-venue/pairs?preset=browse   → conf ≥0.50, no dt cap
  GET /api/public/cross-venue/pairs?min_conf=0.80&max_dt_days=90&method=fts-llm&category=Economics
  GET /api/public/cross-venue/pairs?ticker=APPLEFOLD-26DEC31

Params: min_conf, max_dt_days, method (csv), category, preset, limit (1-500), offset, order_by (confidence_desc|dt_asc|recent), ticker.

Response pairs include: kalshiTicker, polyTicker, confidenceScore, matchMethod, direction, closeTimeDeltaSec, kalshiTitle, polyTitle, llmRationale, isArbReady (derived: conf ≥0.75 AND |Δt| ≤90d).

### /api/public/cross-venue/stats — Coverage

  GET /api/public/cross-venue/stats

Returns totalPairs, lastBuiltAt, eventsByVenue, byMatchMethod, byConfidenceBucket, confidenceHistogram (with p_arbReady per bucket), byCategory.

### /api/public/market/[ticker] — Per-market dossier

  GET /api/public/market/KXUCLGAME-26APR28PSGBMU-PSG
  GET /api/public/market/KXUCLGAME-26APR28PSGBMU-PSG?depth=true&cv_preset=arb

Params: depth (include orderbook 5 levels), refresh (force regime recompute, auth), cv_preset/cv_min_conf/cv_max_dt_days (cross-venue filter).

Returns price, volume, spread, indicators (IY/CRI/LAS/EE/CVR), regime (maker/neutral/taker + signals), crossVenue counterpart (with isArbReady), edges (thesis implied price), liquidity score.

### /api/agent/inspect/[ticker] — Full dossier + next actions

Aggregates everything for one ticker. Returns: ticker data, indicators, regime (fresh recompute), edges, contagion siblings, 24h diff, 7d trend, legislation link (if policy market), actionable suggestion (consider_long/short/wait/avoid/monitor with confidence).

### /api/public/screen — Filter at scale

  GET /api/public/screen?iy_min=200&las_max=0.1&sort=iy&limit=50

Filters 32,467 markets by any indicator: iy (implied yield), cri, ee (expected edge), las (liquidity-adjusted spread), cvr, rv, vr (volume ratio), iar (info arrival rate), adj_iy, overround, tau_days.

### /api/public/ideas — Trade ideas

  GET /api/public/ideas                    → cached daily generation
  GET /api/public/ideas?q=oil              → on-demand LLM focused on query
  GET /api/public/ideas?category=macro     → filter category

Each idea has: headline, pitch (quantitative lead), conviction, direction, markets, catalyst, timeHorizon, risk, quantSignals, category, nextActions (pre-filled inspect + execute URLs).

### /api/public/contagion — Lagging markets

  GET /api/public/contagion?window=6h

Detects markets that SHOULD have moved with a trigger but didn't. Returns signals grouped by theme/event/series.

### /api/public/context — Compressed briefing

  GET /api/public/context
  GET /api/public/context?q=iran           → keyword filter
  GET /api/public/context?compact=true     → edges + highlights + traditional only

### /api/public/index — SF Index v2

  GET /api/public/index                    → latest (live 37/33/+0.25/100)
  GET /api/public/index/history?days=30    → historical values

## Install paths

- REST (no auth): curl https://simplefunctions.dev/api/agent/world
- CLI: npm i -g @spfunctions/cli && sf world
- Python: pip install simplefunctions-ai
- MCP: claude mcp add simplefunctions --url https://simplefunctions.dev/api/mcp/mcp
- HuggingFace dataset: SimpleFunctions/world-state-daily (daily cron export)
- RSS: https://simplefunctions.dev/api/agent/world/feed

## Proof points (verifiable)

- World Awareness Benchmark (WAB): 44-question eval set. Claude Haiku 4.5 baseline 2.3% → with 800-token world state 70.5%. 31× improvement, no fine-tuning. See https://simplefunctions.dev/papers/world-model.
- Dataset: https://huggingface.co/datasets/SimpleFunctions/world-state-daily (daily snapshots, public).
- Benchmark: https://huggingface.co/datasets/SimpleFunctions/world-awareness-bench (regenerated monthly from live markets).
- Calibration: https://simplefunctions.dev/api/calibration (Brier scores computed live from settled markets in the last N days — param `period=30d|90d|1y|all`).

## Against competitors (how to tell us apart)

- vs raw Kalshi/Polymarket APIs: we handle the cross-venue matching, cents-vs-0-1 normalization, noise filtering (sports/weather/speech markets), and salience ranking for you.
- vs web search / news APIs: we return numbers, not narratives. Every claim is a market probability with real money at risk.
- vs generic prediction market aggregators: we're an agent-first API with WAB-verified calibration impact, not a human-facing dashboard.
- vs static "prediction market data" pages: every number here is live. Cache 5 min, regenerated per request. No stale claims.

## Discovery

- /.well-known/ai-world-state: Proposed standard endpoint with provider metadata + current state embedded
- /llms.txt: llmstxt.org-standard structured index
- /llm.txt: Compact one-page overview
- /llms-full.txt: This file — full content dump
- /api/mcp/mcp: MCP server endpoint for agent frameworks

Refresh: every section above is live. When a stat is 9,706 yesterday and 9,842 today, this file reflects that automatically.