Where our data comes from
Crypto prices, market caps and metadata come from CoinGecko’s free API, with CoinPaprika as a documented fallback. Sparklines and OHLC charts come from Binance’s public API. The Fear & Greed Index comes from Alternative.me. Each source is cached for a short interval — from about a minute for tickers and sentiment to longer for charts and metadata — and we advertise refresh cadence honestly on each surface.
When data is missing
If a provider is unavailable we fall back to the last cached value with an “Updated X min ago” note. If we have no value we trust, we render nothing — never a zero, a dash or a placeholder presented as live.
The prediction model
Our price predictions are deterministic. For each asset we generate bear, base and bull scenarios across eight fixed horizons — 24 hours, 7 days, 30 days, 90 days, 1 year, and the years 2027, 2028 and 2030 — from four inputs: price history, realised volatility, liquidity and trading volume, and the prevailing market regime. There is no hand-tuning: the same inputs always produce the same scenarios. The complete model, including the formulas and how each input is weighted, is documented at our methodology engine reference and matches the live behaviour you see on the site.
Model-based scenarios. Not financial advice. Predictions are illustrative projections, not promises or recommendations. Learn how to read them in our guide to reading a prediction.
Frequently asked
Is the market data real-time?
No. Figures are cached for short periods to stay fast and respect provider limits, so they can lag the live market by seconds to minutes.
Are your price predictions financial advice?
No. They are model-based scenarios, not advice. Every prediction carries the line "Model-based scenarios. Not financial advice." and a timestamp.
What happens when a data provider is down?
We fall back to the last cached value with an "Updated X min ago" note, and if we have nothing we trust we show nothing rather than a placeholder.