Portfolio

Portfolio answers "how is this account doing", not "how did this trade do". Every other read in Trade Activity is scoped to one object: a fill, an execution, a position. These two endpoints aggregate all of them.

Use Closed Position when you want one trade. Use these when you want the book.

Both endpoints need two headers:

Neither endpoint pages. There is no limit, no cursor, and no nextCursor. Each returns the whole aggregate in one response. Neither accepts connector either; the venue comes from the trade account.

Every number on this page is a JSON number, not a decimal string. Positions, fills, and closed trades return decimals as strings such as "+41.72". Portfolio returns 41.72. If you feed both into the same parser, portfolio is the one that breaks it. A JSON number is a double, so a very large volume loses precision. Treat these figures as reporting values, not as accounting inputs.

GET /portfolio

curl "https://pro-gateway.pearprotocol.io/portfolio?startDate=2026-01-01&endDate=2026-01-15" \
  -H "x-api-key: $PEAR_API_KEY" \
  -H "x-trade-account-id: $TRADE_ACCOUNT_ID"

{
  "intervals": null,
  "overall": {
    "totalWinningTradesCount": 38,
    "totalLosingTradesCount": 21,
    "totalWinningUsd": 4210.77,
    "totalLosingUsd": 1830.42,
    "currentOpenInterest": 8420.15,
    "currentTotalVolume": 1284500.5,
    "unrealizedPnl": -112.38,
    "realizedPnl": 2380.35,
    "totalFees": 318.94,
    "totalTrades": 59
  }
}

Query parameters

Both dates accept an ISO 8601 string or an epoch-millisecond number. An unparseable date is not a 400. It is silently dropped, and the request behaves as if you never sent it.

overall

Two things to read carefully:

• currentOpenInterest and unrealizedPnl ignore the date window. They are a live snapshot taken when the request arrives. The other eight fields are window-scoped. A query for last March still returns today's open book.
• totalWinningTradesCount + totalLosingTradesCount can be less than totalTrades. A trade that closed at exactly zero net PnL is neither a win nor a loss, but it is still a trade.

intervals

intervals holds prebuilt chart buckets so a dashboard can draw five sparklines from one request.

intervals is null whenever startDate or endDate resolves to a real date. This is not "null when there is nothing to bucket". A custom window returns overall for that exact range and no buckets at all, even on a busy account. Only a bare GET /portfolio — or one whose dates were unparseable and therefore dropped — returns buckets.

When present, intervals carries five arrays of identical buckets:

Every bucket has the same shape:

{
  "periodStart": "2026-01-15T09:00:00.000Z",
  "periodEnd": "2026-01-15T10:00:00.000Z",
  "volume": 12500,
  "openInterest": 8420.15,
  "winningTradesCount": 1,
  "winningTradesUsd": 42.5,
  "losingTradesCount": 0,
  "losingTradesUsd": 0
}

periodStart is inclusive and periodEnd is exclusive. volume is the fill notional in the bucket, and the trade counts and sums follow the same rules as overall.

openInterest is the same value in every bucket. It is today's open interest repeated, not the open interest as it stood in that period. Do not plot it as a time series.

GET /portfolio/analytics

curl "https://pro-gateway.pearprotocol.io/portfolio/analytics?interval=1m&rollingWindowDays=14" \
  -H "x-api-key: $PEAR_API_KEY" \
  -H "x-trade-account-id: $TRADE_ACCOUNT_ID"

{
  "config": {
    "rollingWindowDays": 14,
    "startDate": "2025-12-17T00:00:00.000Z",
    "endDate": "2026-01-15T11:04:22.881Z",
    "fundingSynced": true
  },
  "portfolio": {
    "realizedOnly": {
      "pnlAbsolute": 2380.35,
      "pnlPercent": 0.0412,
      "hitRate": 0.6441,
      "profitFactor": 2.3005,
      "avgWinSize": 110.81,
      "avgLossSize": -87.16,
      "avgReturnPerWin": 0.0231,
      "avgReturnPerLoss": -0.0164,
      "maxDrawdown": -0.1842,
      "sharpeRatio": 1.62,
      "sortinoRatio": 2.08
    },
    "realizedPlusUnrealized": {
      "pnlAbsolute": 2267.97,
      "pnlPercent": 0.0374,
      "hitRate": 0.6441,
      "profitFactor": 2.3005,
      "avgWinSize": 110.81,
      "avgLossSize": -87.16,
      "avgReturnPerWin": 0.0231,
      "avgReturnPerLoss": -0.0164,
      "maxDrawdown": null,
      "sharpeRatio": null,
      "sortinoRatio": null
    },
    "realizedHitRate": 0.6441,
    "profitFactor": 2.3005,
    "avgWinSize": 110.81,
    "avgLossSize": -87.16,
    "breakevenHitRate": 0.4403,
    "cushion": 0.2038
  },
  "assets": [
    {
      "asset": "0",
      "realizedOnly": {
        "pnlAbsolute": 1620.4,
        "pnlPercent": 0.0521,
        "hitRate": 0.6552,
        "profitFactor": 2.51,
        "avgWinSize": 96.4,
        "avgLossSize": -71.2,
        "avgReturnPerWin": 0.0248,
        "avgReturnPerLoss": -0.0177,
        "maxDrawdown": null,
        "sharpeRatio": null,
        "sortinoRatio": null
      },
      "realizedPlusUnrealized": {
        "pnlAbsolute": 1508.02,
        "pnlPercent": 0.0446,
        "hitRate": 0.6552,
        "profitFactor": 2.51,
        "avgWinSize": 96.4,
        "avgLossSize": -71.2,
        "avgReturnPerWin": 0.0248,
        "avgReturnPerLoss": -0.0177,
        "maxDrawdown": null,
        "sharpeRatio": null,
        "sortinoRatio": null
      },
      "unrealizedReturnPerDay": -9.36
    }
  ],
  "pairs": [
    {
      "key": "L:0|S:1",
      "longAssets": ["0"],
      "shortAssets": ["1"],
      "trades": 24,
      "wins": 16,
      "losses": 8,
      "hitRate": 0.6667,
      "avgWinSize": 128.4,
      "avgLossSize": -94.2,
      "expectancy": 54.2,
      "totalPnl": 1300.8,
      "bookPercent": 0.5465
    }
  ],
  "series": [
    {
      "date": "2026-01-14",
      "trades": 3,
      "wins": 2,
      "losses": 1,
      "dayPnl": 84.2,
      "dayVolume": 42500,
      "dayFunding": -3.12,
      "dayNetPnl": 81.08,
      "hitRate": 0.6667,
      "winLossRatio": 2.14,
      "rollingHitRate": 0.6129,
      "rollingWinLossRatio": 1.87,
      "dailyReturn": 0.0084,
      "cumulativeReturn": 0.0398,
      "equity": 1.0398,
      "drawdown": 0,
      "rollingSharpe": 1.44,
      "rollingSortino": 1.91,
      "rollingVolatility": 0.0112
    },
    {
      "date": "2026-01-15",
      "trades": 0,
      "wins": 0,
      "losses": 0,
      "dayPnl": 0,
      "dayVolume": 0,
      "dayFunding": -1.04,
      "dayNetPnl": -1.04,
      "hitRate": null,
      "winLossRatio": null,
      "rollingHitRate": 0.6,
      "rollingWinLossRatio": 1.82,
      "dailyReturn": 0,
      "cumulativeReturn": 0.0398,
      "equity": 1.0398,
      "drawdown": 0,
      "rollingSharpe": 1.39,
      "rollingSortino": 1.86,
      "rollingVolatility": 0.0109
    }
  ]
}

The real response is longer. assets holds one entry per instrument you traded or still hold, pairs one per distinct basket, and series one per UTC day in the window — 30 points for the request above. Only the first entry of each is shown here.

asset, longAssets, and shortAssets hold instrument IDs, so "0" is BTC and "1" is ETH on Hyperliquid. See Instrument ID.

Query parameters

interval never appears in the response. It only picks a start date, and startDate overrides it:

Two behaviours worth planning for:

• rollingWindowDays is clamped to 7–365, not rejected. rollingWindowDays=1 becomes 7, and rollingWindowDays=9999 becomes 365. Read config.rollingWindowDays back to learn what was applied. A non-integer such as 7.5 is a 400.
• The series is on by default. A bare GET /portfolio/analytics returns a year of daily points. Send includeSeries=false when you only want the ratios; it also skips the volume and funding reads, so it is measurably faster.

config

config reports the window the server actually used, after defaults and clamping.

fundingSynced: false means every dayFunding is unknown, not zero. Pear has no funding record for this account, so each dayFunding reads 0 and each dayNetPnl equals its dayPnl — by default, not by measurement. A chart that plots that as "no funding paid" is wrong. Check the flag before you draw funding. Funding reaches Pear through the same exchange sync that settles fills, so an account that has never been synced has no funding to report — see Fills. The flag says only that funding once arrived. It makes no claim about how recent it is.

portfolio

portfolio carries the same metrics under two lenses.

Both variants have the same eleven fields:

maxDrawdown, sharpeRatio, and sortinoRatio are populated in portfolio.realizedOnly and nowhere else. They need a daily return history, and only closed trades have one. In portfolio.realizedPlusUnrealized and in every entry of assets, all three are always null. That is the design, not missing data.

A ratio is null, never 0, when it is undefined. A null sortinoRatio means there was no losing day, or fewer than two days of data — not that downside risk was zero.

portfolio also carries four fields outside the two variants:

In the example, a cushion of 0.2038 says the account wins 20 percentage points more often than its win-to-loss size ratio requires.

assets

One entry per instrument that appears in a closed trade or in an open position, sorted by instrument ID.

Per-asset figures split a basket's PnL back onto its legs. The pairs breakdown below keeps each basket whole. Read Synthetic Position for why the two views differ.

pairs

One entry per distinct basket shape, sorted by absolute totalPnl, largest first. It answers "which pair actually makes me money".

bookPercent does not behave like a percentage of a whole. The denominator is net realized PnL, so a losing basket gets a negative share and a single winning basket can exceed 1 when other baskets lost money. The shares do not sum to 1 in any useful way. Use it to rank contribution, not to fill a pie chart.

series

One point per UTC calendar day from config.startDate to config.endDate, inclusive. Days with no activity are still emitted, so the array length is fixed by the window, not by how much you traded. A year window returns 366 points.

Every rolling field is null for the first rollingWindowDays - 1 points, because the window has not filled yet. Chart them from the first non-null point rather than from the start of the series.

Trades are booked on the day they close, and funding on the day it accrues. A trade held for a week contributes to one dayPnl and to seven dayFunding values.

The equity curve compounds trade returns, not your wallet balance. Deposits and withdrawals do not move it. That is deliberate: it keeps the risk metrics comparable across accounts of different sizes.

What gets rejected

Three inputs are accepted and quietly ignored rather than rejected: an unparseable startDate or endDate, an interval sent alongside a startDate, and any parameter the endpoint does not define. A rollingWindowDays outside 7–365 is clamped, not refused. config is the only reliable record of what the server used, so read it back instead of assuming your parameters applied.

See Error Handling for the two error body shapes.