Read-only market and on-chain data for Bitcoin, Ethereum and more; on-chain from our own nodes.
Scored across 18 tools
Most tools target distinct resources, but there is a genuine overlap cluster: get_asset_metrics and get_metrics_list perform the same lookup (descriptions even say 'the same rows'), and fetch_metric vs fetch_bulk_metrics vs run_formula all read/derive series and could be confused. get_price_history, get_eth_btc_ratio and get_support_levels overlap somewhat with fetch_metric on price data, though the descriptions work hard to differentiate them.
All names are snake_case with a verb-first pattern (fetch_*, get_*, find_*, screen_*, run_*, correlate_*), which is predictable and readable. The one deviation is the get_* family mixing lookup (get_asset_metrics), metadata (get_metric_metadata) and snapshot (get_derivatives_snapshot) semantics, but that's minor.
18 tools is at the heavy end of the typical range for a data-analytics server. Each maps to a distinct CryptoKeyhole feature, but the metrics-lookup pair plus fetch/fetch_bulk duplication suggests the surface could be consolidated to roughly 14-15.
The surface is broad and coherent for an on-chain/metrics data platform: discovery (assets, metrics, screener fields), reads (single, bulk, metadata), analysis (correlate, formula, extremes, screen), and specialized derivatives/ETF/reserves/ratio/support tools. Gaps are minor: no listing of available formulas or categories directly, and no explicit support/health or pagination helper beyond the per-call notes.
How closely two metrics move together over a window: the Pearson correlation of their period-over-period changes (log changes when both stay above zero) and of their levels, and the rolling correlation of changes, from the two exact series. For example BTC's price against its exchange balance, or ETH's price against BTC's.
Inputs: a (required), b (required), end, start, +2 more.
Read several metric selections in one call, each by its exact id. Each item keeps its own provenance and revision; this is not one atomic snapshot. Unsupported selections are explicit per-item errors. The combined result is bounded at 200000 observations and the response-byte limit.
Inputs: requests (required).
Read one metric's observations by its exact id from get_asset_metrics: exact decimal strings with original source clocks, revision, finality and gaps. End is exclusive. No new observation is acquired. Without an API key it serves metrics whose plan is free at 24h, 1w, 1M or 1Y; a chart above the free plan answers with the plan it needs, its free sample window and the free metrics closest to it. A long selection answers with its latest observations and the end to send for earlier ones. For a coin's price on a date, get_price_history answers directly.
Inputs: metric (required), end, sma, asset, +4 more.
Which charts of a coin, or of every coin, stand furthest from normal right now: each chart's latest reading against its own last four years (z-score and percentile), high or low, with the days since it was last this far out. The quickest way to see what is unusual on chain, in derivatives or in flows today. Without an API key it shows the top rows of the free charts and counts the rest by the plan that shows them.
Inputs: asset, limit, query, interval, +1 more.
Search or list the metrics (charts) of one asset by words in query (for example "puell" or "exchange balance") or by category. Use it to find the exact metric id before get_metric_metadata or fetch_metric; ids are looked up here, never guessed. Each row has the plan its data needs; fetch_metric reads rows whose plan is free.
Inputs: asset (required), limit, query, cursor, +1 more.
List the coins CryptoKeyhole covers, each with its code and name. Use a code as the asset argument of the other tools.
Inputs: no named parameters.
The derivatives picture for Bitcoin or Ethereum now: futures liquidations in the last 24 hours (longs and shorts, hour by hour, and the total for every coin covered), the funding rate averaged across venues by open interest, total open interest with its 24-hour change, and the next options expiry's max pain.
Inputs: asset.
Daily net flows of the US spot ETFs of one coin (or of every coin together with ALL), from the issuers' own daily files: the latest day, the largest fund that day, the last 5 and 30 trading days and total holdings.
Inputs: days, asset.
Ether priced in bitcoin: the ETH/BTC ratio now with its 30-day change, 200-day average and all-time high close, and its change between two UTC days when dates are given, from CryptoKeyhole's own hourly price indexes (daily closes since August 2015).
Inputs: end_date, start_date.
Coins CryptoKeyhole can see on chain at each exchange that publishes its addresses (proof of reserves): bitcoin, ether and stablecoins in US dollars, net flows over 24 hours and 7 days, and open flags such as unusual outflows or paused withdrawals. Leave exchange out for every covered exchange.
Inputs: exchange.
Where leveraged Bitcoin or Ethereum futures positions would be liquidated right now: CryptoKeyhole's live liquidation map, estimated by 0.5% price band from 15% below to 15% above the price, with the largest long and short clusters, the totals within 5% and the map's measured accuracy. Redrawn every five minutes. Covers BTC and ETH.
Inputs: asset.
Read one metric's definition by its exact id from get_asset_metrics: what it measures, method, unit, intervals, coverage, source and limitations.
Inputs: metric (required), asset.
Search or list the metrics (charts) of one asset, BTC by default, by words in query or by category. The same rows as get_asset_metrics: each has the exact metric id, name, unit, intervals, history and the plan its data needs; fetch_metric reads rows whose plan is free.
Inputs: asset, limit, query, cursor, +1 more.
A coin's daily closing price on a UTC day, or its change between two days with the highest and lowest daily close in between, from CryptoKeyhole's price history (Bitcoin from July 2010, Ethereum from August 2015, and the other coins covered). Leave end_date out for the change from start_date to today.
Inputs: asset (required), start_date (required), end_date.
List the fields screen_assets can test, with how many coins have each: ids such as market.Mvrv, market.PriceUsdClose or derivatives.FuturesFundingRatePerpetual. Search them by words.
Inputs: limit, query.
Where Bitcoin's price stands against the levels traders watch: the 200-day, 200-week and 1-year averages, the bull market support band, the Pi Cycle Bottom line, the realized price and the power-law lines, each with its distance from the price.
Inputs: no named parameters.
Combine up to eight metrics in a formula and read the result as a series, as CryptoKeyhole's Formulae workbench does: + - * / ^ and functions such as sma, ema, zscore, pct_change, change, corr, rsi, drawdown, realized_vol, rebase, cumsum and ifthen. Name each input with a letter: inputs {"a": "ckh.MVRVPriced", "b": "ckh.SpotPrice"} and expression "zscore(a, 365)". Without an API key the inputs are metrics whose plan is free, at 24h, 1w or 1M.
Inputs: inputs (required), expression (required), end, start, +1 more.
Screen every coin CryptoKeyhole covers by the latest readings of its charts, for example coins with MVRV below 1, funding below zero or price 60% under its high. A condition is field:measure op number, such as market.Mvrv:v<1 or market.PriceUsdClose:p30>10, combined with all(...) or any(...); measures: v the reading, p1 p7 p30 p90 p365 the change in %, d1 ... d365 the change in the chart's unit, z the four-year z-score, q the four-year percentile. get_screener_fields gives the field ids. Without an API key: the first 10 coins at the daily close.
Inputs: dir, sort, where, assets, +2 more.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Read-only Bitcoin blockchain, mempool, mining, market, and on-chain analytics; no API key.
Raw and normalized on-chain data: balances, transactions, blocks and Ethereum validator activity.
Crypto & RWA microstructure, liquidity, flow and raw data via Parquet; live metrics via WebSocket.
Query cryptocurrency on-chain data, OHLCV prices, market data, and Research & QuickTake insights.
Add one secure layer between your agents and this server.
