Free Crypto Datasets
36 datasets on this page carry a recorded last row, and 24 of them gained rows in the 24 hours up to the newest, options-flow-daily.
Every dataset on this page is built from history we accumulate ourselves: the platform records the market around the clock and publishes the raw rows here. Download them, analyze them and build on them for personal and research use with attribution; commercial use comes with the Business plan. The daily series gain one row per UTC day and started collecting on 2026-07-23, so their depth grows every day. Listing scans run hourly and funding history covers a rolling 90-day window. Downloads from a browser need nothing; scripts send a free account key from 2026-09-10.
Citable: The releases frozen up to 2026-09-10 are archived on Zenodo under one concept DOI; the mirrors are no longer updated and the live data stays on this site. doi.org/10.5281/zenodo.22049575
Daily market history
BTC and ETH dominance, total market cap, stablecoin market cap and the Fear & Greed index, one row per UTC day, recorded from our own daily snapshots of public sources, credited in the response header and the JSON meta.
Deeper data: Bitcoin Cycle Indicator Crossover Events 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Market Profile daily
One row per UTC day and perpetual from our own one-minute bars of the Binance USDT-M perpetuals: open, high, low and close, the point of control (poc), the value area low and high (val, vah: 70% of the TPO marks), the initial balance (ib_low, ib_high: the first hour UTC), the volume point of control (vpoc), the TPO mark total, the periods and the minutes the day had. 30-minute TPO periods, buckets of bucket_pct percent of the day's open. Columns: date, symbol, exchange, bucket_pct, open, high, low, close, poc, vah, val, ib_high, ib_low, vpoc, tpo_total, periods, minutes.
A day with fewer than 20 hours of minute bars has no row. Naked points of control are derived when the API is read, not stored.
Deeper data: Market Profile Daily History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
TradFi weekend and overnight gaps
One row per window, underlying and checkpoint for the stock, index and commodity perpetuals: the reference close (the median venue price in the last 10-minute snapshot before 16:00 New York), then each weekend checkpoint (sat_12, sun_12, sun_20, pre_open, open) or night checkpoint (utc_04, pre_open, open) with the venue median, the venue count, the spread between venues and the gap to the close in percent. Columns: window, close_day, open_day, symbol, market, checkpoint, target_utc, snapshot_utc, price_median_usd, venues, spread_pct, gap_pct.
No holiday calendar: on a US market holiday the close or the open falls on a shut session. The raw 10-minute venue snapshots are on the paid venue-markets-history route.
Deeper data: Venue Market Snapshots (10-minute) 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Liquidation venue share daily
One row per UTC day and counted liquidation venue: the venue's recorded liquidations in USD, its share of the day's recorded total, the day total and its feed kind (full, sampled, delayed); a venue has rows from its first counted day. Columns: date, exchange, coverage, liquidations_usd, share_pct, day_total_usd, partial_day.
Shadow venues and venues without a public feed are not in the total, so the total is a floor of the market. Hourly rows per venue are on the paid venue-share-history route.
Deeper data: Liquidation Venue Share Hourly History 0.1 USDC per call · Raw Liquidation Events 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Desk notes
One row per note the desk published: publish time in UTC, day, slug, card type, kind, the desk's 0 to 10 news value, headline, the English text, the card's data time, the board it reads from, the coins it names, the translation status and the permanent URL. Written by the desk writer (an AI model) from the platform's own recorded derivatives data and checked against it before publishing; translations are on the API and the paid route.
Every note is one observation about the recorded data at the time shown, not advice. The text is the desk's own; a wrong note is hidden with a reason, never rewritten.
Deeper data: Desk Notes Archive 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Tokenized stocks daily
One row per UTC day, issuer, wrapper and chain for the onchain stock wrappers of xStocks, Robinhood Stock Tokens, Coinbase Tokenized Stocks, Binance bStocks, Ondo Stocks, Backpack Securities, Reality (Bitget), Superstate Opening Bell, Figure, Dinari dShares and Securitize (Exodus) and the exchange pairs (chain cex): the day's close price, the reference share price, the premium in percent, 24h volume and liquidity of admitted pools, and the supply read after midnight (shares, dollars, holders, multiplier), then price_source: the price that valued the dollars (reference, issuer_oracle for the issuer's own price, wrapper for its own pool or exchange price, empty when the row has no dollars). xStocks carry their circulating supply under chain all.
Premiums are measured against a fresh reference in an open session only; a closed session leaves the premium empty.
Deeper data: Tokenized Stocks Hourly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Exchange insurance funds daily
One row per UTC day, venue and asset for the insurance funds Binance, Bybit, OKX, Gate, MEXC, Bitfinex, Crypto.com, Delta Exchange, dYdX and Paradex publish: the day's last balance in the asset, its USD value (dollar stables 1:1, other assets at our recorded price, empty when unpriced) and the number of pools. OKX files its published venue-wide total as asset USD; Gate's rows before 2026-09-23 come from Gate's own published history.
Read every hour from each venue's public endpoint; a fund is a balance the venue reports, not an audit of its reserves.
Deeper data: Exchange Insurance Fund Hourly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Jupiter Perps daily
One row per UTC day, market and metric at 00:00 UTC for Jupiter Perps on Solana: exact long and short open interest from the on-chain custody state, long and short pool utilization, hourly borrow rates, custody AUM, price and 24h volume for SOL, ETH and BTC, plus JLP pool AUM, price and APR.
Read from the official Jupiter perps API every 10 minutes and recorded hourly on our side; open interest is program state, not an estimate. The daily file keeps the midnight row, the paid route the full hourly history.
Deeper data: Jupiter Perps Hourly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Solana Perps daily
One row per UTC day, venue, market and metric at 00:00 UTC from the Solana Perps board (Jupiter, Pacifica, Phoenix, GM Trade, Velocity and Bullet; Jupiter's rows live in the Jupiter Perps daily file): oi_usd, oi_long_usd and oi_short_usd (long and short on the pool venues), volume_24h_usd, funding_hourly_pct and mark_usd. Columns: date, venue, market, metric, value, source.
Open interest is one-sided on every venue, the convention exchanges and DefiLlama use: an order book's contracts times mark, a pool venue's long plus short halved.
Deeper data: Solana Perps Hourly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Daily symbol board
Daily funding rate, open interest and price rows for the major-USDT-perpetual basket: one row per symbol per UTC day.
Append ?symbol=BTCUSDT to filter the download to a single contract.
Deeper data: Open Interest 5m History 0.1 USDC per call · Funding Rate History 0.1 USDC per call · Venue Market Snapshots (10-minute) 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Perpetual listing events
Every USDT-perpetual listing and delisting event our hourly scanner has recorded across Binance, OKX, Bybit, Gate.io, HTX and BingX.
Rows with listing_known=0 were already trading when tracking began, so their true listing dates are unknown; only contracts first seen after the baseline carry real listing dates.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Borrow rates
Hourly margin borrow rates per venue and asset (USDT, USDC, BTC, ETH and major alts on every venue whose borrow history we can read), as published plus an annualised figure.
Read from each venue's borrow history endpoint; Binance publishes a daily rate, OKX an annual one, both stored as published with apr_pct alongside. More venues join with read-only keys.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Trading fee schedules
Current base tier maker and taker per venue and market type, one row per schedule or per pair where the venue prices pairs individually.
Read from each venue's fee endpoint; a changed schedule closes the old row and opens a new one, so fee changes on the page are derived events. Venues with authenticated fee endpoints join with read-only keys.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Settlement prices
Settlement prices recorded as dated futures deliver on the venues we record (Bybit, Crypto.com, Deribit, HTX, OKX today; more as their endpoints answer), one row per contract and settlement.
Read from each venue's settlement history endpoint for the dated futures on file; options settle by the thousand and have their own recorders. The expiry calendar itself comes from the daily market lists of 53 market list venues.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Withdrawal networks
Current deposit and withdrawal network table on every venue we record (20+ exchanges): per asset and network the deposit flag, withdrawal flag, withdrawal fee and minimum, one row per pair.
Recorded from each venue's public currency list; a flag or fee change closes the old row and opens a new one, so suspensions and resumptions on the page are derived events. Venues with authenticated currency lists join once a read-only key is configured.
Deeper data: Withdrawal Networks History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Leverage tiers
Current leverage ladder for every recorded perpetual on every venue whose ladders we read: tier, notional floor and cap, maximum leverage and maintenance margin rate, one row per tier.
Recorded from each venue's risk limit endpoint; a changed ladder closes the old rows and opens new ones, so the change log on the page is derived, never typed. Binance joins once a read-only key is configured.
Deeper data: Leverage Tiers History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Daily liquidations
Daily long and short liquidation totals in USD per symbol and exchange, recorded from our own live streams across Binance, Bybit, OKX, Gate, HTX, dYdX, Aster, Lighter, Kraken Futures and GMX, plus Deribit's delayed flag. One row per UTC day, symbol and exchange.
Recorded events, not estimates: Binance throttles its public stream heavily, so totals are a floor. History accumulates daily from the collection start; no earlier data exists.
Deeper data: Raw Liquidation Events 0.1 USDC per call · Bulk Export: Liquidation Events 15 USDC per month · Commercial use: Business plan · redistribution of derived data: Scale
Largest single liquidations, daily top 10
The ten biggest single liquidation prints of every finalized UTC day across the counted venues: rank, symbol, venue, side (SELL = a long was liquidated, BUY = a short), price, quantity, notional and millisecond event time.
Ranks come from our own recorded prints, so a venue that throttles its public stream (Binance: one print per second per symbol) can only rank lower, never higher. Aggregated liquidations on one venue and symbol within the same millisecond count as one print.
Deeper data: Largest Liquidations History 0.1 USDC per call · Raw Liquidation Events 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
RSI heatmap daily
One RSI(14) reading per UTC day, contract and interval (15m, 1h, 4h, 12h, 1d, 3d, 1w, 1M) for the top 400 Binance crypto perpetuals and every TradFi perpetual: the 00:00 UTC row of our hourly record.
History exists only from the day recording started. RSI is a level, not a signal; the file is for testing what the level meant on this market.
Deeper data: RSI Heatmap Hourly History 0.1 USDC per call · Bulk Export: Whale Prints, Positioning, RSI 10 USDC per month · Commercial use: Business plan · redistribution of derived data: Scale
TradFi perpetuals daily
One row per UTC day and Binance TradFi perpetual (US, HK, KR and CN stocks, commodities, indices): closing open interest, the day's settled funding, and liquidations recorded on every venue we collect from.
Open interest and funding history exist only from the day recording started; earlier rows do not exist. Liquidations are recorded prints, a floor.
Deeper data: TradFi Perpetuals Open Interest 5m History 0.1 USDC per call · TradFi Perpetuals Funding History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Kimchi and Turkey premium daily
One row per UTC day, venue and coin: the day's last 10-minute price on Upbit, Bithumb, Coinone (KRW at the ECB rate) and BtcTurk (USDT) versus Binance spot at the same bucket, as a percent premium, for BTC, ETH and XRP.
Both legs are read at the same 10-minute bucket, so the premium is a same-moment gap, not a daily close difference. KRW conversion uses the ECB reference rate captured at collection time.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Hyperliquid whales daily
One row per UTC day and coin: long and short notional, long share, position counts and the largest single position on each side across the 1000 largest Hyperliquid accounts by equity (the 300 largest before 2026-09-30), from the day's last 5-minute scan.
Addresses are not in this file; the paid event route carries per-address changes. History exists only from the day scanning started.
Deeper data: Hyperliquid Whale Position Events 0.1 USDC per call · Bulk Export: Whale Prints, Positioning, RSI 10 USDC per month · Commercial use: Business plan · redistribution of derived data: Scale
Hyperliquid liquidation price levels
The current liquidation price map of every coin the 1000 largest Hyperliquid accounts by equity hold: one row per coin, side and band of 0.25% of the mark, with the band price, the signed distance from the mark, notional and position count, from the latest 5-minute scan.
A snapshot, not a series: each download is one scan, named in the as_of column. Positions without a liquidation price are not placed; the hourly position archive is the paid history route.
Deeper data: Hyperliquid Positions Hourly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Positioning daily
One row per UTC day, contract, exchange and metric at 00:00 UTC: Binance global and top-trader long/short ratios and taker buy/sell ratio, Bybit share of accounts long, OKX long/short ratio, top-trader position ratio and taker buy share, Gate, HTX and Bitget account and position ratios (from 2026-09-20), for the 30 most traded perpetuals; plus Bitfinex margin long and short position size in coins and the long share on its spot margin book (from 2026-09-22; margin borrowers, not perpetual accounts).
Exchange-published aggregates recorded hourly on our side; the daily file keeps the midnight row, the paid route the full hourly history.
Deeper data: Positioning Hourly History 0.1 USDC per call · Bulk Export: Whale Prints, Positioning, RSI 10 USDC per month · Commercial use: Business plan · redistribution of derived data: Scale
Spot order book depth daily
One row per UTC day, coin and venue at 00:00 UTC: mid price, bids and asks within 2% (USD), book reach and level count, plus the full 0.1% bucket arrays (50 per side within 5% of mid) as JSON, for BTC, ETH, SOL, XRP, DOGE, ADA, LINK, AVAX, LTC, BNB on Binance, Coinbase, Kraken, Bybit, OKX, Gate, MEXC, HTX, Bitget, Bitstamp, Gemini, Bitfinex, BingX, Crypto.com, Backpack, Binance US, Poloniex, LBank, Bitrue, XT, HashKey, DigiFinex, Bullish, WEEX, Deribit, Aster, Bitvavo, Independent Reserve and BtcTurk.
REST books reach different depths per venue (Coinbase the whole span, Binance about 1.5%, Kraken 2%, Bybit and OKX 0.3%); buckets beyond a venue's reach are unknown, not zero.
Deeper data: Spot Order Book Depth History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
US spot ETF daily flows
Daily net inflows, total net assets, cumulative inflow and value traded for the US spot Bitcoin, Ethereum, Solana and XRP ETFs, in USD. One row per US trading day and asset (SOL since 2025-10-28, XRP since 2026-09-08).
Recorded daily by ByKaranteli. Finalized US trading days only; the current day stays provisional until every fund has reported. XRP rows start on 2026-09-08 and each carries the cumulative inflow since launch.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
US spot ETF flows by fund
Daily net inflow, net assets, cumulative inflow and management fee for every US spot Bitcoin, Ethereum, Solana and XRP ETF (IBIT, FBTC, GBTC, ARKB, ETHA, BSOL, XRPC and the rest), in USD. One row per US trading day and fund.
Recorded daily by ByKaranteli since 2026-07-30; a fund gets a row for a day once it has reported that day. History grows from the collection start.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
ETF issuer primary-source series (IBIT, XRPC)
Daily shares outstanding, NAV and fund-level net assets read straight from the issuers' own pages (BlackRock for IBIT, Canary for XRPC), plus the issuer-implied flow (day-over-day change in shares x NAV) and the recorded per-fund flow and net assets for the same date, so the issuer's own numbers and the daily fund figures can be compared row by row.
BlackRock states IBIT shares as of the previous day's position, so the IBIT implied flow can trail the trade-date fund figure by a day. Canary states XRPC shares by trading day, so the XRPC implied flow lines up with the same day's fund figure.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
CME futures positioning (COT)
Weekly CFTC Commitments of Traders positioning for CME Bitcoin, Micro Bitcoin, Ether and Micro Ether futures: open interest plus long and short contracts for every trader category.
Official CFTC regulatory data, republished as filed. Positions as of each Tuesday, published Friday.
Deeper data: CFTC COT Weekly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Coinbase Premium daily history
Daily closing price gap between the US-regulated venue and the global market price for BTC and ETH, in percent, with both closes included. One row per day and symbol.
Positive premium = US buying pressure. Both closes are UTC daily candles; the current day is excluded until final.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Implied volatility index (DVOL) daily
Daily closes of Deribit's 30-day implied volatility index (DVOL) for Bitcoin and Ethereum options. One row per day and currency.
The crypto equivalent of the VIX, from listed option prices.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Options chain daily snapshot
The latest daily open-interest snapshot of the listed BTC and ETH option chains on Deribit, Bybit, Binance, OKX and Delta Exchange: per-instrument open interest in coins, mark implied volatility, underlying price and the venue.
Serves the most recent chain; the accumulating history powers our GEX and term-structure charts.
Deeper data: Options OI By Strike History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Options chain hourly
Every listed BTC and ETH option on Deribit, Bybit, Binance, OKX and Delta Exchange, captured every hour: open interest in coins, mark implied volatility, underlying price and 24h volume per venue, plus Deribit's mark, bid and ask in coins. Columns: h, venue, currency, instrument, expiry, strike, type, open_interest, mark_iv, mark_price, bid_price, ask_price, underlying_usd, volume_usd.
The file serves a window; the paid route serves every recorded hour. Quotes are Deribit's: the other venues quote in USD, USDT or USDC and those columns stay empty.
Deeper data: Options Chain Hourly History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Metric context daily spine
One daily value per metric across the whole platform: BTC close, Coinbase premium, DVOL, ETF net flow, Fear & Greed, daily funding and VPIN, plus quantum exposure. The substrate behind every "what happened after days like this" table.
Rows are labeled live (recorded same day) or backfill (reconstructed from a dated series).
Deeper data: Funding Rate History 0.1 USDC per call · Open Interest 5m History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Narrative theme indices daily
Eight equal-weight crypto narrative indices (AI, RWA, DePIN, meme, L1, L2, DeFi, quantum) from Binance daily closes, rebased to 100 on 2025-01-01, one row per theme per day.
Baskets are fixed lists in the codebase and change only with a dated commit.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Venue markets daily
Per venue, market and symbol: the last snapshot of each UTC day with price in USD, 24h volume, open interest and 8h-equivalent funding for every perpetual, spot and TradFi perpetual feed we poll.
The 10-minute rows, recorded since 2026-08-25, are the paid venue-markets-history route; this daily cut is free.
Deeper data: Venue Market Snapshots (10-minute) 0.1 USDC per call · Bulk Export: OI, Funding, Venue Snapshots 10 USDC per month · Commercial use: Business plan · redistribution of derived data: Scale
Whale tape daily
Per symbol per day: number of $1M+ aggressive prints, buy and sell notional and the buy share, from our own trade sockets.
Raw prints are the paid whale-trades-raw route.
Deeper data: Whale Tape Raw Prints 0.1 USDC per call · Bulk Export: Whale Prints, Positioning, RSI 10 USDC per month · Commercial use: Business plan · redistribution of derived data: Scale
Stablecoin pegs daily
USDT, USDC, DAI and FDUSD per day: how many spot venues quote them against USD and the min, median and max price at the day's last snapshot.
A venue-level print outside the band is not a redemption figure; the median is the number to watch.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Bitcoin quantum exposure daily
Daily first-party measurement of quantum-exposed Bitcoin classified from our own node's UTXO snapshot, with base_height, base_hash and txoutset_hash on every row so any figure can be re-verified against any Bitcoin node.
The watched dormant-P2PK outpoint set has its own CSV at /api/public/quantum/watchset.csv. Erratum: launch day 2026-08-13 was measured twice; the duplicate row (height 962221) was removed from this series on 2026-08-27 and the canonical row is the first completed measurement (height 962212). Archives sealed before that date still carry both rows.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Basis and carry daily
Quarterly futures basis snapshots per symbol and tenor: spot, future and the annualized cash-and-carry yield, captured daily. The trade behind much of the hedge-fund positioning in the COT report.
Small but append-only; every row is a dated capture, never recomputed.
Deeper data: Funding Rate History 0.1 USDC per call · Commercial use: Business plan · redistribution of derived data: Scale
Execution cost ladder hourly
Hourly order-book walks per major perpetual: the average slippage in bps for $10K, $100K, $1M and $5M market orders on both sides, plus visible book depth. Order books cannot be backfilled; this archive is why the history exists at all.
Fees excluded by design: this is pure visible-book impact.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Options tape daily rollup
Daily aggregate of the recorded Deribit and OKX options tape: trade count, notional and premium volume, put share and block-trade count per currency.
The raw tick tape stays internal; this rollup is the honest public series.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
Funding rate settlement history
Settled funding rate history per symbol: a 90-day settlement table with averages, extremes, positive share and hour-of-day seasonality, backfilled from Binance settlement data.
One page per pilot contract: /symbols/<SYMBOL>/funding. Every Binance USDT-M contract's settlements are in the paid funding-history route and the bulk-derivatives file.
Deeper data: Commercial use: Business plan · redistribution of derived data: Scale
License and attribution
The data we compute on this page is published under the ByKaranteli Data Licence: free for personal use, research, dashboards, models and articles, with a visible attribution to ByKaranteli (bykaranteli.com). Commercial use needs the Business plan and redistribution of derived data needs Scale (see /developers#tiers); those rights cover the data we compute. A series that carries a provider's or a publisher's values stays under that source's terms and is credited where those terms ask for it, and a single venue's raw records are for your own use (Terms of Service, "Data, licences and attribution"). Files published before 2026-09-10 on the GitHub, Kaggle, Hugging Face and Zenodo mirrors were dedicated to the public domain under CC0 for our own computations; those mirrors are no longer updated. If you republish derived numbers, include the dataset's date range.
The files published up to 2026-09-10 stay as a frozen archive on GitHub (bykarantelicom/crypto-datasets).
Free weekly funding report
One email every Monday: the week's funding extremes, biggest OI shifts and the dominance move. No spam, unsubscribe in one click.
Double opt-in. No tracking pixels. Unsubscribe link in every email.
Questions
Where do these datasets come from?
Is there a citable archive with a DOI?
Which licence covers the datasets?
Source: ByKaranteli, https://bykaranteli.com/data, data as of 2026-10-23 08:13 UTC