# HYPO · LLM crawler manifest

> For agents needing the full reference (JSON shapes, URL patterns, stable
> contracts, statistical foundations), see `llms-full.txt`.


> HYPO.markets is a sovereign autonomous live market mirror.
> It re-publishes Hyperliquid perpetual futures, Hyperliquid HIP-3 prediction
> markets, and Polymarket binary outcomes with full statistical analysis on
> every page. Data is fetched per request, not stored.

## Site

- Operator: HYPO · Aurelius Foundation
- Domain: https://hypo.markets
- Licence: CC0 / public domain for all derived statistics
- Storage policy: no persistence; every page fetches the upstream feed live
- Attestation: every dashboard publishes a SHA-256 of the source snapshot

## Endpoints

### JSON / data

- `GET /api/feed` — combined Hyperliquid + Polymarket snapshot
- `GET /api/feed/hyperliquid` — Hyperliquid perps + predictions snapshot
- `GET /api/feed/polymarket` — Polymarket markets snapshot
- `GET /api/asset/[asset]` — per-asset snapshot + 24h history (live; no cache)
- `GET /api/asset/[asset]/stats` — pre-computed statistics JSON (no OHLC, just metrics)
- `GET /api/stats?assets=a,b,c` — batch stats for explicit asset list (max 50)
- `GET /api/stats?venue=pm&max=20` — top-N curated stats per venue
- `GET /api/asset/[asset]/stream` — Server-Sent Events live stream (SSE)
- `GET /api/og/[asset]` — dynamic 1200×630 PNG OG card with sparkline + verdict
- `GET /api/changelog` — machine-readable ship log (mirrors /changelog)
- `POST /api/model/[asset]/q` — push external model probability (body: `{q, sigma?, source?, ttlSeconds?}`). Auth: `Authorization: Bearer <HYPO_MODEL_API_TOKEN>`. Rate limit: 30/hour/IP. GET reads current (no auth); DELETE clears (auth required).
- `GET /api/book/[asset]` — L2 order book for pm-<slug> assets (asks + bids per YES/NO side). ISR 4s.
- `POST /api/archive/snapshot` — write today's snapshot to disk; sweep stale archives (auth + 6/hr rate-limit). GET lists available dates.
- `GET /api/related/[asset]` — same-topic basket via TF-IDF cluster; `?eg=1` adds Engle-Granger cointegration screen against basket peers.
- `GET /api/health` — liveness + upstream age

### HTML / dashboards

- `GET /` — live market mirror landing page
- `GET /feed/hl-{ticker}` — Hyperliquid perpetual dashboard (e.g. `/feed/hl-btc`)
- `GET /feed/hl-pred-{slug}` — Hyperliquid HIP-3 prediction market dashboard
- `GET /feed/pm-{slug}` — Polymarket prediction market dashboard
- `GET /sim/{asset}` — interactive NOSTRADAMUS Position Analytics simulator (URL state: ?price=&q=&side=&kf=&bank=&ttl=&qsd=&fee=&spread=&entry=&seed=)
- `GET /sim/portfolio` — multi-position portfolio simulator (correlation, correlated-Kelly haircut, book-walk slippage)
- `GET /markets` — sortable filterable table of every tracked market across all venues
- `GET /compare-by-category` — top movers leaderboard across 7 topic areas
- `GET /arb` — adaptive multi-route arbitrage strategy classifier
- `GET /methodology` — statistical foundations reference (Kelly, Brier, Murphy, GARCH, ADF, KPSS, VR, Engle-Granger, Hurst, MC)
- `GET /status` — live operations dashboard (snapshot ages, SSE fan-out)
- `GET /embed/{asset}` — iframe-friendly mini card (60s auto-refresh)
- `GET /archive/{YYYY-MM-DD}` — historical daily snapshot
- `GET /compare/{asset-a}/vs/{asset-b}` — pairwise comparison
- `GET /related/pm-{slug}` — same-topic basket (TF-IDF cluster) + Engle-Granger cointegration screen
- `GET /positions` — local positions ledger (browser-only localStorage; mark-to-market against snapshot)
- `GET /archive/resolved` — Polymarket markets whose endDate has passed, with ex-post calibration table
- `GET /compare/radar?assets=a,b,c` — N-asset radar chart on 6 normalised quant axes (max 6 assets)
- `GET /backtest` — strategy classifier ex-post replay on resolved PM markets (per-strategy realised return aggregate)
- `GET /category/{slug}` — single-category landing page

### Feeds / discovery

- `GET /sitemap.xml` — full URL inventory including every asset detail page
- `GET /robots.txt` — crawl policy
- `GET /feed.xml` — RSS 2.0 of top movers per polling cycle

## Dashboard semantics

Every per-asset detail page renders, in order:

1. Big H1 + plain-English summary block (auto-generated from current data)
2. Live overlay (poll every 8s) with animated hero numbers, sparkline, donut, bars
3. §1 24h time-series (line + candles + volume bars)
4. §2 returns distribution (histogram + Q-Q plot for normality)
5. §3 sample moments (μ, σ, median, quartiles, skew, kurtosis, CI)
6. §4 returns analytics (Sharpe, Sortino, Calmar, annualised vol)
7. §5 risk metrics (VaR₉₅, VaR₉₉, ES₉₅, max drawdown)
8. §6 time-series structure (autocorrelation, Hurst exponent, OLS trend t-stat)
9. §7 microstructure / liquidity
10. §8 position sizing (Kelly, half-Kelly, quarter-Kelly)
11. §10+ academic addon: hourly heat row, equity + underwater drawdown,
    rolling-window Sharpe/vol/ρ, hypothesis tests (Jarque-Bera, Ljung-Box,
    Dickey-Fuller, runs), DFT periodogram
12. §∞ provenance + SHA-256 attestation

## Schema.org coverage

Pages embed:

- `FinancialProduct` or `Quotation` with statistical PropertyValues
- `Dataset` (with `variableMeasured`, `temporalCoverage`, `distribution`,
  `license: CC0`) — surfaces in Google Dataset Search
- `FAQPage` (prediction markets only; question = FAQ entry)
- `Event` (prediction markets with a known endDate)

## Citation

If you cite HYPO data in research or downstream analytics:

  HYPO.markets — Live market mirror with statistical analysis.
  Aurelius Foundation, Próspera ZEDE. https://hypo.markets
  Snapshot SHA-256: {attestation hash visible on each page}

CC0 — no permission required, attribution appreciated.

## Contact

ai@cireasa.org

## Notes for crawlers

- All pages are server-rendered (SSR) on every request — content visible to
  crawlers exactly as to users
- No JavaScript required to read the data
- 24h history is fetched live from Hyperliquid candleSnapshot and Polymarket
  CLOB price-history endpoints; no client-side persistence
- Page-rendered timestamp + SHA-256 attestation make every page reproducible
- The site supports many locales; per-asset detail pages are language-agnostic
- robots.txt allows full indexing; sitemap covers all asset pages