Questions, answered plainly

EdgeDepth answers one question: what usually happened after moments like this one?

We record 660+ Binance USDT-M crypto and TradFi perpetuals, plus every Hyperliquid perpetual, continuously. You can watch them live in a browser tab, rewind any recent moment and replay it tick by tick, or search the last 90 days of recorded history in plain English and get every time your setup occurred, what followed each time, and a replay of any of them.

It counts what happened. It never forecasts what happens next. The live view is free; the recorded history is what a subscription buys. Start with the record, or see what costs money.

Anything you can describe with the readings we store. Every recorded minute carries 33 searchable features: order-flow pressure, book tilt, funding, open interest, liquidations, volatility, candle shapes and more.

Combine a few into a condition (“open interest surges, then $5M is liquidated within the hour”) and the search returns every minute that matched, what followed each one, and a replay of any of them. You can also point at a minute instead of describing it: the reader tells you what was true then and turns it into a search you can edit before anything runs.

The one thing you cannot do is filter on what happened afterwards. What you can ask lists the whole vocabulary, and you can run one yourself.

Yes. The terminal is C++ compiled to WebAssembly and draws through WebGL2, so it runs in an ordinary browser tab: no install, no download, and the free tier needs no card.

The terminal covers what it does, and the terminal at a glance walks through the panels.

The live view is free. The recorded history is what you pay for.

Free is the full live terminal across 660+ Binance USDT-M crypto and TradFi perpetuals, plus every Hyperliquid perpetual, replay of one fully archived 24-hour UTC day on 6 majors, archived events from the last 7 days (plus any event we mark public, whatever its age), free lessons, and sample searches. It needs an account, never a card.

Pro adds the rolling 30-day replay and lookup window across 660+ recorded Binance USDT-M crypto and TradFi perpetuals, every archived event, the full lesson catalog, the advanced chart layers (footprint, volume profile, liquidation tiers), alerts, paper trading in replay, and 10 record searches a month.

Research extends replay and lookup to 90 days, raises the allowance to 300 searches a month, and adds the full research workflow: multi-condition and sequence searches across every recorded market, outcome comparisons, saved searches and live alerts, reports and CSV export, and the same engine from code or an AI agent.

Pricing has the line-by-line comparison.

Yes. The terminal client is open source under AGPL-3.0, and the MIT-licensed gateway runs it on Binance’s free public market data with one command. Self-hosted you get the chart with real history, DOM ladder, orderbook, trade tape, liquidations, the candle-derived liquidation Field, VPVR, TPO, paper trading and 1s / 5s / 15s / 30s candles built from the raw trade stream. No account, no API key, no time limit.

What needs the hosted product is everything computed from the recorded history: VPIN, positioning, the modelled liquidation heatmap and Levels, patterns, scanner scores, contagion, deep historical replay and the courses taught inside the terminal. The open source page has the full split and the one-command quick start.

No. Every reading is named in plain words first and its technical id second, and the numbers are always counts: how many times something happened, out of how many chances it had. You can read a result without knowing how the reading behind it is computed.

When you do want the background, each idea has a short guide: order flow, VPIN, book imbalance, funding, open interest and liquidations. The rest are here.

No, and the honest version matters, because there is a conditions library on this site and alerts really do ping you.

What a signal group sells: someone else’s rule, usually secret and never counted, delivered as a trade call with an implied promise about what happens next.

What an EdgeDepth alert is: a condition you can read, counted against the recorded history before it was armed. When it happens again, the ping says exactly one thing: the thing you counted just occurred. No direction, no entry, no exit, no prediction. The history rides along with it: how often it happened before, what followed each time, and a replay of every one. You cannot arm an alert for a condition nobody has run.

So the ping is real. The promise usually attached to it is what we refuse to sell. What you do when it fires is your call, on your own account. How live alerts work.

No. Every count shows how many times something happened, out of every chance it had: the wins, the losses, and the times nothing followed.

You also cannot ask us to cherry-pick. A search like “show me the times this fired and price then went up” is refused by design, because filtering on what happened afterwards is how backtests lie. You see the full set first, then what followed across all of it.

Base rates explains the number most trading stats leave out, and selection bias explains the trap.

We say so. If a measurement was not available at some moment (a market gap, a value still warming up), that moment is left out of the count and the exclusion is reported with the result. A gap is never quietly counted as a zero.

The same rule runs everywhere: the Hyperliquid liquidation layer states how much of the market it actually covers, estimates are labeled as estimates, and a search that finds nothing says “this did not happen in the window we scanned” rather than pretending the question was never asked. The record explains what sits behind the counts.

So anyone can check it. Every result carries a short ID that pins exactly what was asked, which frozen code measured every value, and which state of the recorded history was scanned.

Run it again next week: same answer. Send the link to a friend: same answer. If the record has grown since, the ID says so instead of quietly changing the number. Our figures are meant to be checkable, not just quotable, and every match still opens in tick-by-tick replay so you can watch it for yourself. Determinism, explained.

Conditions is a shelf of 26 curated conditions: named market states such as an extreme funding rate, a one-sided tape, or open interest accelerating into liquidations.

Each one has a page per market showing how often it happened there, what usually followed it with the count beside every rate, and a replay link for each occurrence. It is a starting shelf, not a recommendation list: a condition that fires constantly and leads nowhere reads exactly like that, which is the point of counting it. Base rates explains how to read those numbers.

For the markets we publish, yes. Every market page shows what that market’s readings look like at the most recent minute in the record, which of the tracked states are true right now, and what usually followed each of them, with the count beside every rate.

The chart underneath covers the last month: the candles, the moments that fired (each one replayable), and the liquidation layers. Every reading also carries its percentile, so an ordinary value is easy to tell apart from an unusual one. Percentiles explains that chip.

One page per UTC day. The daily brief lists what fired across the tracked markets, and for each one what usually followed the previous times it fired on that market, with the count beside every rate and a replay link to the moment itself.

It is assembled from the same counts as the rest of the site: no commentary, no opinions, nothing written by a model. Read it on the site, or leave your email and get it once a day. The typical-worst-drawdown figure beside each median is explained in MFE and MAE.

A scan returns every matching minute, what followed each one, and a replay of any of them. Use it when you want the examples.

A base rate answers how common the condition was across everything scanned, denominator attached. Use it when you want the yardstick.

A comparison puts the matches beside every comparable minute where the condition did not hold, which is the only way to see whether the setup changes anything at all.

Fixed-anchor stratification is a fourth tool, built and audited but not released yet to the browser, the API or the agent tools. It takes one already-defined set of matches and splits it by a second condition that was true, false or unavailable at that same minute, so the set and its timestamps cannot drift between the groups the way they do when you run two searches and subtract one from the other.

How research works covers the workflow, and reading a distribution covers the results.

Yes. The same engine answers over a public REST API and a read-only MCP server, so Claude, Cursor, Codex and other agents can call it as grounded tools: list the readings, turn plain English into a proposed search, count occurrences, page through them, read a market’s state at a past minute, compare a matched set against every comparable minute, and fetch published reports.

The agent has to stay inside the same fixed vocabulary you do, and every answer keeps its denominators and its reproducibility key. It cannot place a trade, connect an exchange account, arm or change your alerts, or invent a field that does not exist.

Developers is the overview; the API and MCP docs have the full reference.

The recorded market state, not a chart of it: individual trades, full-depth order-book updates, the DOM ladder, liquidity heatmaps, liquidations, open interest, the analytical views the terminal draws from them, and every received best-bid and best-ask price and quantity change, preserved per symbol.

Playback is deterministic, so the same recorded input rebuilds the same market sequence every time. Hyperliquid history is still being backfilled; it becomes searchable and replayable once it passes the same completeness checks as the Binance record. Starting a replay shows how to open one.

Candle replay redraws bars from historical prices. EdgeDepth rebuilds the market sequence that produced those bars: the individual trades, the order-book changes as they arrived, the liquidity, the liquidations and the open interest.

You are not watching a recording of a chart. You are watching the market that drew it. Replay has the detail.

Replay: Pro covers a rolling 30-day window across the recorded Binance markets. Research extends that window to 90 days, and Free includes a smaller allowance on the majors. Named archived events stay replayable beyond those windows under their stated free, public or paid access rules.

Search: Pro’s included searches look back 30 days. Research looks back 90 days and raises the monthly search allowance. Every result states which markets and dates it scanned, and anything that was missing, so you never have to guess what an answer covers. Newly recorded history becomes searchable only after it passes our completeness checks.

No. EdgeDepth already records 660+ Binance USDT-M crypto and TradFi perpetuals continuously, so you can return to a timestamp or an archived event even if you were nowhere near a screen at the time. That is the whole idea: the market gets recorded whether or not anyone was watching it.

Every Hyperliquid perpetual is live in the terminal. Hyperliquid history is still being backfilled; it becomes searchable and replayable once it passes the same completeness checks as the Binance record.

Live coverage spans 660+ Binance USDT-M crypto and TradFi perpetuals, plus every Hyperliquid perpetual, down to every received best-bid and best-ask price and quantity change, preserved per symbol.

Search and public tick replay currently cover 660+ recorded Binance USDT-M crypto and TradFi perpetuals: crypto majors, long-tail alts, and USDT-settled perpetuals tracking Nvidia, Tesla, AMD, QQQ, gold, silver and oil. Those are Binance derivatives, not shares or direct exchange feeds, and Hyperliquid contracts follow Hyperliquid’s own venue and settlement rules.

Hyperliquid history is still being backfilled; it becomes searchable and replayable once it passes the same completeness checks as the Binance record. Cross-exchange research is in the works, and results will always say which venue they came from before any Binance-to-Hyperliquid comparison is allowed. The data page covers what we store and for how long.

Yes. Guided lessons can pause before a decisive moment, ask you to read the market, and let you place a paper trade before the recorded outcome is revealed. No money moves, and the market you are reading actually happened.

No. The free and public lessons cost nothing, and the full catalog is included with Pro: there are no per-lesson purchases.

Interactive questions and paper trading in replay work today. Saved quiz scores and progress across the catalog are still to come. Lessons shows how one runs.

No. EdgeDepth is market data and education. Nothing on this site or in the product is a recommendation to buy or sell anything, and no result is promised. Perpetual futures are leveraged instruments and carry a substantial risk of loss.

Counts describe what the record contains. They are observed history, never a forecast, and what you do with them is your decision. The disclaimer is the formal version.

Downtime-credit terms are not finalized yet, so we are not going to imply that one exists. Any credit policy will be published once the billing behavior and commercial terms are approved. Until then, the terms are the only commitment we have made.