Hyperliquid userFills limit: get more than 10,000 fills

Hyperliquid's userFills returns at most 2,000 fills, and userFillsByTimeonly reaches a wallet's 10,000 most recent. Older fills are gone from the API. Ironflow's GET /v1/fills?address= pages through every fill a wallet made since 25 May 2025, 1,000 per page.

API

GET https://api.ironflow.sh/v1/fills?address=0x... no key

SQL

SELECT ... FROM hl.wallet_fills WHERE address = '0x...'

MCP

get_fills on https://mcp.ironflow.sh/mcp

The short version#

• Hyperliquid: 2,000 fills per call and the 10,000 most recent per wallet.
• Every fill: /v1/fills?address=, oldest first, with a cursor. History starts on 25 May 2025.
• Who needs it: 221 wallets made more than 10,000 fills in the 24 hours to 12 November 2026.

Hyperliquid's limits#

From the info endpoint page. Each of these also costs weight 20 plus more for every 20 fills returned, out of 1,200 per minute per IP (rate limits).

Who hits 10,000#

In the 24 hours to 21:10 UTC on 12 November 2026, 95,256 wallets traded on Hyperliquid. 825 made more than 2,000 fills and 221 more than 10,000. The busiest made about 597,000. For those 221, userFillsByTime cannot reach back even one day.

Every fill over REST#

import os
import time
from datetime import datetime, timezone

import requests

ADDRESS = "0xYOUR_ADDRESS"
KEY = os.environ.get("IRONFLOW_KEY")  # optional
headers = {"Authorization": f"Bearer {KEY}"} if KEY else {}
params = {"address": ADDRESS, "from": "-24h", "limit": 1000}  # "-14d" with a free key

fills = []
while True:
    r = requests.get("https://api.ironflow.sh/v1/fills", params=params, headers=headers, timeout=30)
    if r.status_code == 429:
        time.sleep(int(r.headers.get("Retry-After", 5)))
        continue
    r.raise_for_status()
    body = r.json()
    fills += body["data"]
    if not body["pagination"]["has_more"]:
        break
    params["cursor"] = body["pagination"]["next_cursor"]


def utc(ms):
    return datetime.fromtimestamp(ms / 1000, timezone.utc).strftime("%Y-%m-%d %H:%M")


print(len(fills), "fills,", utc(fills[0]["timestamp"]), "to", utc(fills[-1]["timestamp"]), "UTC")

Keyless, for one of those wallets on 12 November 2026:

10606 fills, 2026-11-11 21:50 to 2026-11-12 21:30 UTC

• With an address there is no cap on the window per request, only your plan's history.
• Each fill has price, size, side, dir, fee, realized_pnl, oid, tid and hash.
• Add market=BTC-PERP for one market. Endpoint reference: list fills.

SQL#

hl.wallet_fills holds the last 92 days sorted by wallet, so one address reads fast. Fills per hour:

curl -X POST https://api.ironflow.sh/v1/query \
  -H "Content-Type: application/json" \
  -d @- <<'EOF'
{"sql": "SELECT toStartOfHour(ts) AS hour, count() AS fills FROM hl.wallet_fills WHERE address = '0xYOUR_ADDRESS' AND ts > now() - INTERVAL 1 DAY GROUP BY hour ORDER BY hour DESC LIMIT 3"}
EOF

For totals over months use hl.wallet_daily. Tables and limits are in the SQL guide.

How far back by plan#

The other route is Hyperliquid's S3 archive: node_fills_by_block has every fill of every wallet, in requester-pays buckets you download and filter yourself.

Questions#

Related guides#

• Hyperliquid trades and fills data: API, archive and history by wallet
• Hyperliquid historical liquidations API
• Hyperliquid data sources compared
• Hyperliquid SQL: query every fill, liquidation and transfer

References#

• Hyperliquid docs: Info endpoint. userFills, userFillsByTime and userTwapSliceFills limits.
• Hyperliquid docs: Rate limits. Request weights per IP.
• Hyperliquid docs: Historical data. node_fills_by_block on S3.