Get Hl Liquidation Bars
GET/v1/cryptodata/backtesting/hl-liquidation-bars
Incremental liquidation flow per closed bucket, summed from the exact Hyperliquid fill tape — the burst clock ``/backtesting/liquidations``' rolling-24h level cannot give you. Each row is one UTC clock-aligned ``interval`` bucket: ``long_usd`` / ``short_usd`` (notional liquidated on each side), ``total_usd``, ``n_long`` / ``n_short`` (fills) and ``max_fill_usd``. Buckets with no fills are not returned. ``start`` and ``end`` are floored to the grid, so the bucket containing ``now`` is never served — its fills are still forming and live on ``/backtesting/hl-liquidations``. Omit ``coin`` for one market-wide row per bucket (``coin`` null, window ≤ 31 days); give a coin or comma list for per-coin rows (window ≤ 93 days). Every bar is auditable: the ``audit`` block says how to pull the raw fills that summed to it. Same tier as the tape (Pro, Pro Plus included).
Price
$0.01/ callEdge cache
Off
Uptime 24h
100.00%
p95
2.11s
Parameters
| Name | In | Type | Description |
|---|---|---|---|
| coin | query | string | One HL coin name (BTC, kPEPE) or a comma list (up to 25) for per-coin bars. Omit for ONE market-wide series summed over every perp. |
| interval | query | string = 15m | Bucket width: 5m, 15m or 1h (UTC clock-aligned) |
| start* | query | string | Start time (ISO 8601 or unix ms); floored to the interval |
| end | query | string | End time (ISO 8601 or unix ms, defaults to now); floored to the interval, EXCLUSIVE |
| bounds | query | string = half-open | Interval convention for `start`/`end`. `half-open` (default) is [start, end) — `end` is EXCLUSIVE, so consecutive windows tile without repeating a boundary row. `closed` is [start, end] — `end` INCLUSIVE, the pre-2026-08-08 behaviour; use it only to keep an existing integration byte-identical. |
| limit | query | integer = 1000 | |
| cursor | query | string | Opaque page cursor — the `next_cursor` of the previous response, verbatim. Resumes strictly after the last row served, keyed on the table's full primary key, so a page boundary inside a same-millisecond burst loses nothing (unlike `start = last.time + 1`). `start` / `end` still bound the window. |
| format | query | string = json | json or csv. CSV has no envelope: paging state travels in the `X-Has-More` / `X-Next-Cursor` response headers. |
Code
# 1. No key, no payment: the gateway answers 402 with the price
curl -i "https://api.realtimeapi.ai/v1/cryptodata/backtesting/hl-liquidation-bars?start=2026-10-28"
# 2. Pay: retry with a signed x402 payment (any x402 client builds this header)
curl "https://api.realtimeapi.ai/v1/cryptodata/backtesting/hl-liquidation-bars?start=2026-10-28" -H "PAYMENT-SIGNATURE: $X402_PAYMENT"// npm i @x402/fetch @x402/core @x402/evm viem
import { wrapFetchWithPayment } from "@x402/fetch";
import { x402Client } from "@x402/core/client";
import { ExactEvmScheme } from "@x402/evm/exact/client";
import { privateKeyToAccount } from "viem/accounts";
const client = x402Client.fromConfig({ schemes: [{ network: "eip155:8453", client: new ExactEvmScheme(privateKeyToAccount(PRIVATE_KEY)) }] });
const pay = wrapFetchWithPayment(fetch, client);
const data = await (await pay("https://api.realtimeapi.ai/v1/cryptodata/backtesting/hl-liquidation-bars?start=2026-10-28")).json();
console.log(data);# pip install x402 eth-account requests
from eth_account import Account
from x402.clients.requests import x402_requests
session = x402_requests(Account.from_key(PRIVATE_KEY))
print(session.get("https://api.realtimeapi.ai/v1/cryptodata/backtesting/hl-liquidation-bars?start=2026-10-28").json())curl "https://api.realtimeapi.ai/v1/cryptodata/backtesting/hl-liquidation-bars?start=2026-10-28" -H "X-API-Key: $RTAPI_KEY"Frequently asked questions
How do I get get hl liquidation bars from CryptoDataAPI?
Send GET https://api.realtimeapi.ai/v1/cryptodata/backtesting/hl-liquidation-bars. Pay $0.01 per request with x402, or send an X-API-Key header.
What does GET /backtesting/hl-liquidation-bars cost?
$0.01 per successful request in USDC on Base. Failed requests are never charged.
Can I call this endpoint from an MCP client?
Yes. Connect any MCP client to the RealtimeAPI server and use call_feed with this endpoint's URL, or get_quote to check the price first.