Futures 4 venues
- BinanceUSD-M / COIN-M
- Bybitlinear / inverse
- OKXswap
- Hyperliquidperp
Meridian normalizes crypto market data across venues into reliable, historical, provenance-aware datasets for quants, agents, and research systems.
01 / COVERAGE
One schema across venues. Exchange-specific quirks stay out of your research pipeline.
Futures-OI coins in rollout, added as coverage clears provenance checks.
02 / ONE SCHEMA
Query once and get rows back in an identical shape, ready to concat, join, or backtest against.
| exchange | close_price | funding_rate | open_interest | liq_volume | buy_volume |
|---|---|---|---|---|---|
| binance | 67,412.30 | +0.0084% | 2,184,902,114 | 4,120,004 | 18,204,113 |
| bybit | 67,409.80 | +0.0079% | 911,204,552 | 1,884,210 | 7,041,209 |
| okx | 67,415.10 | -0.0012% | 604,118,932 | 902,441 | 4,118,204 |
| hyperliquid | 67,410.55 | +0.0102% | 388,204,410 | 1,204,918 | 3,004,112 |
03 / DATA QUALITY
Meridian is built on captured exchange events and disclosed coverage, not interpolation dressed up as certainty.
Every row traces back to a captured exchange event. Nothing is modeled to look complete.
Each series ships with start date, gap log, and venue-specific notes before you query.
Know which endpoint, capture method, and timestamp produced each row.
If an exchange does not publish a field, Meridian says so instead of estimating it.
04 / DEVELOPER EXAMPLE
The SDK returns typed, columnar market data with the cleanup already handled.
from meridian import Client
client = Client(api_key="mrd_live_...")
df = client.futures.bars(
symbol="BTCUSDT",
interval="1h",
exchanges=["binance", "bybit", "okx"],
start="2026-06-01",
)
df.groupby("exchange").mean()
| ts | exchange | close | oi_close | funding |
|---|---|---|---|---|
| 06-01 00:00 | binance | 67,190.20 | 2,101,004,112 | +0.0081% |
| 06-01 01:00 | binance | 67,242.55 | 2,109,441,203 | +0.0079% |
| 06-01 00:00 | bybit | 67,188.90 | 901,204,552 | +0.0074% |
| 06-01 00:00 | okx | 67,201.40 | 598,204,113 | -0.0009% |
05 / CUSTOM BARS
Time bars hide information when activity is uneven. Meridian constructs bars directly from the trade tape.
06 / ACCESS
Pay for the rows you pull. No seat counts, no annual contract, no sales call to get a key.
For prototyping and small research jobs.
For live systems and continuous backtesting.
For funds and infrastructure teams with volume needs.
07 / EVERYTHING CALLABLE
Every icon below is a documented endpoint or SDK surface — see the docs for schemas and examples.
MERIDIAN
Reliable historical datasets for quants, agents, and research systems.