Every crypto trading floor has a CoinGlass tab open somewhere. Liquidation heatmaps, aggregated open interest, funding rates across thirty-odd exchanges: as a way of seeing the derivatives market right now, it has no serious rival, and at $29 a month for entry API access it is among the cheaper habits on a desk. We sell a competing data product, so read what follows with that in mind. The honest version of this comparison, though, is less “which is better” than “these are different instruments that happen to both have REST endpoints.”
What CoinGlass is
CoinGlass grew out of derivatives monitoring, and it shows in the best sense. Coverage is broad, aggregating more than thirty exchanges. The indicator set is the one traders actually watch: open interest, funding, long/short ratios, liquidations, options, ETF flows. Updates land within a minute on paid plans, the v4 API presents everything through standardized schemas, and lately they have begun selling heavier products too, order book snapshots among them. For discretionary trading, risk monitoring, alerting, or market commentary, the subscription earns its keep on day one.
Where the dashboard stops
The friction appears when you try to feed a model. CoinGlass gates granularity and history by tier: on the $29 Hobbyist plan, historical series stop at four-hour resolution; the $79 Startup plan reaches hourly bars, but only 180 days back and nothing finer than thirty minutes; a full year of hourly history, and the licence for commercial use, start at the $299 Standard plan; the $699 Professional tier raises the ceilings again. Daily series run back to 2019 on every plan, which is fine until the effect you care about lives inside the day.
Access is JSON over per-minute quotas: 30 requests a minute at the entry tier, 1,200 at the top paid tier. That shape is right for a dashboard refreshing a chart and wrong for research. Pulling three years of hourly history across a few hundred symbols through a rate-limited JSON API is an afternoon of pagination code and a long-running script, repeated whenever your universe changes.
There is a subtler limit underneath. CoinGlass values are aggregates designed for glancing: open interest blended across venues, an estimated liquidation figure, a ratio. For monitoring, exactly right. Research usually wants the opposite: per-exchange attribution, documented construction, and a guarantee that the value stamped 14:00 was knowable at 14:00, because a backtest silently rewards every violation of that guarantee. Display products rarely make that promise, and nothing on a chart breaks when they do not keep it.
What Aperiodic is
Aperiodic is built for the pipeline rather than the wall monitor. The catalog covers order flow (signed volume split by trade size, with a toxicity score), slippage and market impact at fixed notionals, top-of-book and multi-level order book imbalance and liquidity, intra-bar volatility, VWAP and TWAP, and the derivatives set: funding, basis, open interest, and perpetual prices. Every series is reported per exchange, point-in-time, and validated before publication. Candles come as TrueOHLCV, aggregated from locally timestamped trades, so bar boundaries reflect what was observable rather than what was stamped after the fact.
Delivery is built for bulk: parquet files over a REST API, a Python SDK that hands you dataframes, one-minute resolution on every self-serve tier, and multi-year history on every tier, with full history and 15-second bars on institutional plans. Coverage runs deep rather than wide: Binance futures and OKX perpetuals from 2020, plus Hyperliquid perpetuals, with more venues on the roadmap. If you need options flows, ETF data, or liquidation maps across thirty venues today, keep CoinGlass; we do not attempt breadth we cannot validate.
The pricing coincidence
Billed yearly, Aperiodic's self-serve tiers work out to $29, $79, $299, and $699 a month: the same headline figures as CoinGlass's monthly plans. (CoinGlass discounts annual commitments too, so at equal commitment their effective numbers dip lower.) The symmetry makes the real question unusually clean. At the same price point, one product hands you aggregated indicator endpoints; the other hands you columnar research datasets with per-exchange attribution and a validation trail. Which dollar is better spent depends on whether the numbers end up in front of a human or inside a model.
Side by side
| CoinGlass | Aperiodic | |
|---|---|---|
| Built for | Watching markets in real time | Feeding research pipelines |
| Coverage | 30+ exchanges, aggregated | 3 venues deep, per-exchange attribution (expanding) |
| Data shape | Indicator endpoints (JSON) | Columnar datasets (parquet) |
| Finest historical resolution | 4h at $29; 30m at $79; finer on higher tiers | 1 minute on every tier; 15s/30s on institutional |
| History depth | Daily to 2019 on all plans; intraday gated by tier | Multi-year on all tiers; archives from 2020; full on institutional |
| Bulk access | 30 to 1,200 requests/min depending on tier | Presigned parquet files; one request per month of data |
| Commercial use | From the $299 Standard plan | Included with institutional plans |
| Liquidations, options, ETF flows | Yes, aggregated | No (funding, basis, OI, perp prices per venue) |
| Point-in-time discipline | Display-first aggregates | Local timestamps, validation, documented lineage |
Which one to buy
Choose CoinGlass when a human is the consumer: discretionary trading, risk dashboards, alerts, market commentary, or a fast answer to what positioning is doing right now. Its breadth is real, the charts are good, and the price of admission is low.
Choose Aperiodic when code is the consumer: backtests you intend to trust, ML features, execution cost models, cross-sectional signals. Construction discipline only pays rent there, which is why that is where we spend our effort.
Many teams should simply run both; at these prices the combined line item is not worth a meeting. Put CoinGlass on the wall monitor and Aperiodic in the pipeline, and be suspicious of any workflow where the data flows in the other direction.