HyperPulse Docs

Public methodology for wallet analytics, funding signals, and tomorrow's market bias.

HyperPulse is a Hyperliquid analytics layer built for active traders. It combines public wallet data, fill history, funding history, and market structure to explain what happened in a portfolio and what the tape has historically implied next.

Overview Data Sources Portfolio Analytics Market Signals Tomorrow Bias Wallet Modes Limitations FAQ

Overview

What HyperPulse does

HyperPulse has two jobs. First, it reconstructs a trader's recent behavior from Hyperliquid wallet activity and turns that into portfolio analytics such as win rate, expectancy, drawdown, asset breakdowns, timing patterns, and funding drag. Second, it evaluates market conditions across listed assets and converts raw funding, open interest, and price action into a cleaner signal layer.

Read-only analytics first. HyperPulse can analyze any wallet address without asking for a private key.

Evidence over vibes. Funding and sentiment labels are only meant to be directional when the underlying sample is meaningful.

Public methodology. The docs below describe the current production calculations in plain English.

Inputs

Data sources and source of truth

HyperPulse uses Hyperliquid account state, public fill history, funding history, and market candles as its primary inputs. Hyperliquid remains the source of truth for balances, open positions, fills, and funding payments. HyperPulse layers portfolio reconstruction and market inference on top.

Portfolio inputs

Wallet account state, marked spot balances, open positions, fills, realized fees, and funding cash flows.

Market inputs

Funding history, current funding APR, open interest changes, 24h price change, and recent candle history for forward-return studies.

Portfolio

How portfolio analytics are calculated

Portfolio analytics are built from completed round trips. HyperPulse groups fills by asset and direction, derives weighted average entry prices, and closes trades when the tracked position returns to zero. Funding paid or earned during that trade window is merged into the resulting trade record.

Net P&L

Closed trading P&L + net funding - fees

This is the top-line realized result from completed trades in the selected history window.

Win Rate

Winning round trips / total round trips

Trades are grouped into round trips by asset and direction. Open positions are not counted as wins or losses until closed.

Fees Paid

Sum of fill fees across the selected trade set

Shown separately because fee drag matters for active strategies, especially short-horizon systems.

Equity / Account Value

Perps equity + full marked spot wallet

Perps equity comes from Hyperliquid margin account value. The spot wallet adds idle USDC plus marked non-USDC balances such as HIP-3 spot holdings. Staked HYPE is not included.

Drawdown

Largest peak-to-trough decline on cumulative realized P&L

Used for risk-adjusted metrics such as Calmar and recovery factor.

Expectancy

(Win rate × average win) - (Loss rate × average loss)

Useful for understanding whether the system has a positive edge per completed trade.

Accounting note

HyperPulse currently defines displayed trading equity as perps equity + the full marked spot wallet. Staked HYPE is intentionally excluded from this number so the dashboard stays aligned with immediately usable trading capital.

Signals

How funding signals are produced

HyperPulse does not rely on a raw funding threshold alone. It first measures how extreme the current funding APR is versus roughly 30 days of history, then studies whether historical funding extremes actually correlated with forward price returns over a default 24 hour horizon.

Signal pipeline

1. Convert funding history into annualized APR series.

2. Rank the current funding APR inside the recent distribution.

3. Pair each historical funding point with forward 24h returns from candles.

4. Compute funding/forward-return correlation and compare extreme buckets.

5. Promote to a stronger label only when the relationship is material enough to matter.

Crowded Long / Crowded Short

Only shown when current funding is extreme relative to history and forward returns have shown a material relationship with funding.

Funding Elevated / Funding Cheap

Used when funding is rich or cheap versus recent history, but the correlation with forward returns is too weak to make a stronger claim.

Neutral / Low Confidence

Default state when history is too sparse or the relationship between funding and forward returns is not reliable enough.

Confidence is graded from sample size and absolute correlation strength. Sparse history or weak correlation forces the signal back toward low confidence rather than overstating precision.

Sentiment

How HyperPulse estimates tomorrow's bias

The HyperPulse sentiment model is a composite regime indicator with a short-horizon directional overlay. The headline score runs from fear to greed. The directional overlay estimates next-session bias using BTC as the market anchor.

Headline regime score

40% funding regime, 35% market breadth, 25% volatility breadth. Funding regime itself blends median funding direction, signal bias, and a 24h versus 7d funding moving-average regime check.

Tomorrow bias model

BTC 24h momentum, BTC 48h momentum, BTC open-interest change, and a contrarian funding term are blended into a directional score. Current production weights are 35%, 25%, 25%, and 15%.

This model is intentionally short-horizon. It is designed to frame the next trading session, not to act as a multi-month valuation model.

Access

Wallet modes, privacy, and execution

HyperPulse supports read-only analysis and connected trading workflows. Read-only mode is the safest and simplest path for analytics because it only needs a wallet address. No private key is required to view a public account.

Read-only

Loads account state and public history for any wallet address. No trading permissions. No private key requested.

Browser wallet

Requests a connected wallet, then approves a local Hyperliquid agent key for order execution. Trading permissions stay scoped to the approved agent.

Honest note on Privy

Privy is best treated as an identity and wallet-discovery layer here, not as proof that HyperPulse has found your exact Hyperliquid trading account automatically. If your live trading address differs from the wallet Privy shows first, choose the linked wallet explicitly or paste the exact address you trade on.

Caveats

Known limitations and interpretation rules

HyperPulse is analytics software, not a custody layer. It does not take possession of user funds.

Signals are descriptive and probabilistic, not guaranteed forecasts. A strong historical relationship can still fail in live markets.

Some analytics are intentionally gated behind minimum sample sizes to avoid false precision.

Displayed account value is optimized for trading clarity, not tax or brokerage-style reporting.

Open positions can change faster than the portfolio polling interval, so intraminute UI differences versus Hyperliquid may occur.

FAQ

Frequently asked questions

Why can HyperPulse show a wallet without connecting a browser wallet?

Because wallet analytics are built on public account state and public fill history. Read-only mode only needs the wallet address.

Does Privy email login automatically connect my Hyperliquid trading wallet?

Not always. Privy can authenticate you by email and may also expose an embedded wallet, but your actual Hyperliquid trading account is often a separate linked external address. HyperPulse now lets you explicitly choose which Privy wallet to view or trade from, and you can always paste the exact address you trade on.

Why might HyperPulse differ slightly from Hyperliquid?

The app groups fills into round trips and computes portfolio metrics on top of those trades. Hyperliquid is the source of truth for raw balances, positions, fills, and funding.

Does HyperPulse include staked HYPE in equity?

No. The current production calculation uses perps equity plus the full marked spot wallet. Staked HYPE is excluded so the trading balance stays interpretable.

When are AI Insights hidden?

If the underlying trade sample is too small. HyperPulse suppresses asset, hour, and performance claims when sample thresholds are not met.