HYPERBASIS HyperBasis Terminal Guide API Pricing Swagger Docs

HyperBasis Data Coverage

A complete and honest inventory of what HyperBasis collects, how deep each dataset goes, and — explicitly — what does not exist.

1. Coverage at a glance

Dataset	Instruments	Depth	Cadence
Deviation (Hyperliquid vs TradFi)	98 RWA instruments	90 days at 5-minute granularity (tradable universe >95% covered)	90-second live cycle
Minute-resolution deviation	58 assets	~18 days (growing daily)	aligned per minute
TradFi 1-minute reference	~90 assets	up to 2 years (top assets full)	1-minute bars
Market snapshots (mark price)	full Hyperliquid RWA universe	rolling ~3 weeks	10-second WebSocket snapshots
Funding & open interest	full universe	multi-month snapshots, 365-day backfill in progress	10-minute snapshots
Capacity / depth	full universe	rolling	live
Whale positions	282-address pool	rolling	30-minute scan
Liquidation events	full universe	rolling	real-time

2. The deviation dataset in detail

The flagship dataset records, for every Hyperliquid RWA perpetual, the percentage difference between the Hyperliquid mark price and the TradFi reference price. As of the latest collection the dataset holds more than 2.5 million 90-day-window rows, and the tradable universe averages above 95% coverage against the 90-day, 5-minute standard. The coverage metric is published live on the HyperBasis homepage.

3. Honest gaps

Zero-volume instruments (for example ALUMINIUM, VIX, DXY, CORN, WHEAT, KRW, NIFTY, IBOV, GEV, KSTR): these Hyperliquid perps never traded, so no tradable history exists. HyperBasis shows N/A for them instead of inventing data.
Minute-resolution deviation depth: the Hyperliquid candle API only serves ~3.5 days of 1-minute history, so minute-level deviation cannot be backfilled beyond the local snapshot window. It accumulates organically and is labeled as such.
Same-day TradFi minutes: on free data plans, same-day US equity minute bars lag one trading day. Live intraday reference pricing uses the 90-second quote layer instead.
Borrow costs: Layer A (stock-analysis short-interest proxy + estimated fee) is live; real broker borrow rates (Layer B) and options-implied borrow (Layer C) are not yet wired and are labeled honestly where shown.
Macro calendar: seeded fallback events are used until a TradingEconomics key is provisioned.

4. How coverage is measured

HyperBasis runs a completeness audit against a 90-day window of 5-minute buckets (25,920 expected points per asset). An asset below 95% coverage is flagged INCOMPLETE in the terminal and its leaderboard confidence is reduced, rather than hidden. The audit re-runs continuously and the numbers on this page are regenerated live.

5. Data quality flags

OK — deviation computed from a fresh TradFi quote (≤5 minutes old).
TRADFI_STALE — the TradFi leg is older than 5 minutes; shown but never trusted as a live signal.
anomaly — deviation rows beyond the historical anomaly threshold are flagged and excluded from statistical calculations.

Honest-limitations policy. HyperBasis labels every data gap instead of hiding it. Where a source does not exist (e.g., zero-volume Hyperliquid instruments that never traded), the product shows N/A rather than fabricated values.