Screenshot
Clipboard copy isn't supported on this device/browser. Tap the button below to download instead, or press and hold (or right-click) the image and choose Save Image / Copy Image — then paste it into a message to share it.
Username and password, stored here only — no email, nothing shared with anyone else.
Potential reversal alerts. Fulcrum watches every tracked coin on 15m, 1h, 4h and 1d and flags swing points where several independent conditions line up at once.
I’m busy with life, same as everyone. Fulcrum exists because I couldn’t sit and watch a hundred seventy coins across four timeframes all day — and I didn’t want to keep finding out about the good ones after they were over.
So I leave it open while I work on other things, or when I’m spending time with family. When it alerts me, I go and look.
Once a signal confirms, that is when I pull up the chart and do my own technical analysis on it — and only then do I work out an entry, a stop loss and a take profit. Fulcrum does not give me any of those three. It tells me where to look; the levels I actually trade are mine, off my own chart. If I don’t like what I see when I get there, I ignore it and go back to what I was doing.
That last part matters more than it sounds. Ignoring a signal is a normal, correct outcome, not a missed opportunity. The tab’s job is to spend its attention so you don’t have to spend yours; deciding what to do with what it finds is still entirely yours.
Rare by design. Counting early warnings and the follow-ups that pass all gates, expect roughly twenty-five to thirty entries a day across all coins and all four timeframes — of which about ten pass all gates. Not a stream. If the list looks quiet, that is the intended behaviour — a signal on every swing would carry no information.
A swing high is only a swing high if what comes after it is lower. A pivot cannot be confirmed at the moment it forms — it has to survive a stretch of later trading first, and that stretch is proportionally longer on slower timeframes.
This is not a delay that could be engineered away. Until that trading has happened there is nothing to judge — the information does not exist yet.
That wait means an all-gates-passed signal can arrive well after the pivot itself, longer on slower timeframes. So on every timeframe, Fulcrum speaks twice.
First an early warning, sent as soon as a pivot has formed and survived its first test. It is marked EARLY · UNCONFIRMED and it is exactly that — nothing more. Then a follow-up, either ALL GATES PASSED or DIDN'T HOLD (price went back through the pivot). Both are sent — you are never left with a warning that quietly goes nowhere.
About two in three early warnings go on to pass all gates, and the ones that fail tend to fail quickly. An early warning is by construction less reliable than one that's passed all gates; that is the price of hearing about it sooner. It arrives substantially ahead of the follow-up on every timeframe.
Tap any signal to see its full history — when the pivot formed, when it was flagged, and how it resolved. Nothing is overwritten or removed.
"All gates passed" does not mean the move is guaranteed, and it does not mean price is going up or down 100%. It means every requirement the signal checks was met and the pivot then held afterward — a swing point formed and was not exceeded. That is a description of something that has already happened, not a forecast of what comes next or a promise of what happens after.
The same is true of the direction wording. Possible pivot down refers to the turn that has been observed at that level, not a prediction that price will keep falling. What happens after the signal is not part of the claim.
This describes a signal that has passed all gates. Because a pivot only clears all its gates some time later, there is always a real wait between the pivot forming and the alert arriving — longer on slower timeframes, shorter on faster ones.
A signal is better read as marking a level that mattered than as a moment to act on.
The first line is what most recently happened and when — alerted, early warning, or didn’t hold — and the "ago" always counts from that. The second line is the pivot: the bar the signal marks, and where a chart would draw it. Alongside it you’ll see the bar it confirmed on, when an open warning settles by, or how many bars it survived before the pivot was exceeded. Tap any row for the full history.
Only signals that fire from now on appear here. History is never backfilled into the list, so nothing shows up already hours old as though it had just triggered.
The buttons filter the list. The same coin can turn on one timeframe while doing nothing on another, so every signal carries the one it fired on. Slower timeframes wait longer, fire far more rarely, and give up less of the move.
Think of the fast tiers as scalps and the slow ones as swings. 15m and 1h pivots tend to hold up over a short window, so they suit a quick, tactical read. 4h and 1d mark bigger structure that takes longer to prove out one way or the other — better read as a level that mattered than a moment to act on.
Low-liquidity coins pull the odds down some. Real thin volume, not market cap, is what actually does it — a couple of coins here rank fairly high by market cap but trade so little that they needed a manual exception to even be included, and those specifically read noisier than the rest of the list, especially on the fastest tier. Cards for those coins carry a LOW LIQUIDITY tag next to the ticker so you can see it without having to already know the name. Treat a signal on one of these with extra caution — it isn't disqualifying, but it's a real reason to double-check what you're looking at before acting on it.
On a wide screen, Small Cap coins (rank 121–200) get their own column, pooling their signals from all four timeframes in one place. Every other column — 15M, 1H, 4H, 1D — shows Core coins (top 120 by market cap) only, no small caps mixed in. They're kept apart because 15M Small Cap signals call the real top or bottom noticeably less often than Core ones do. Read a Small Cap signal on any timeframe with the same extra caution as a LOW LIQUIDITY one above.
A coin needs a couple of hundred closed bars behind it before any of this can be measured, so a recently listed one stays quiet on the slower timeframes for a while — it may run on 15m and 1h while 4h and 1d have nothing to say yet. That is missing history, not a missing signal, and it fills in on its own.
Think of Fulcrum as a potential-setup indicator. It flags where conditions have lined up on a chart you probably weren't watching — a suggestion of what to look at, not an instruction to act. In practice that means a signal is the start of your work, not the end of it: it confirms, you open the chart, you do your own analysis, and you decide the entry, the stop loss and the take profit. Fulcrum supplies none of those three numbers. Whether there is a trade in it is a separate question, and it is yours to answer against your own criteria: your timeframe, your risk, the levels you actually care about, whether you want exposure to that coin at all.
Plenty of signals will fire on names you don't trade, at prices you don't like, in conditions you'd rather sit out. Those are still working as intended — they simply aren't your setup. The tab exists to point your attention somewhere specific across a hundred seventy coins and four timeframes. Deciding whether what you find is worth trading is the part it can't do for you, and shouldn't.
When it calls a pivot, it's usually right about the general area — but treat that price as a zone, not an exact level to trade off of. Price often keeps going a little past it before actually turning, or the real turn lands a few ticks earlier or later than the number shown. Give it room.
Tap the ☆ next to a coin on any signal to save it. The ★ WATCHLIST button at the top then shows only signals for your starred coins. Your list is stored only in this browser and is private to you.
Signals on your starred coins also appear in the “Since your last visit” card and the ★ pill on the Heatmap tab (see its i button).
This flags where a turn may be forming. It is not a buy or sell signal, not a prediction, and not financial advice. It reports a condition that has occurred, and says nothing about what price does next. Trade your own decisions.
Tap the 🪲 at the top of the page to report something broken or confusing. Your message goes to the people who run Tephra Terminal. You can add a Discord or email if you want a reply, and before sending you can see exactly which technical details (tab, coin, window size, recent errors) go with it.
A random ID this browser generated for anonymous usage counts — no account, no personal data, and it never leaves this site. Quote it if you report a problem, or send it to the site owner to be whitelisted from rate limits.
…
@scottytradesSS4 (Twitter)
Join Tephra Terminal Discord — questions, bug reports, or to hear when new features land. Members also get free $CRUM tokens.
Potential reversal alerts, watching two separate universes side by side — Stocks (S&P 500 + Nasdaq-100, 518 tickers deduped, 4H/1D/1W) and Forex (7 FX majors, 4H/1D/1W) — flagging pivotal areas where several independent conditions line up at once. Same detector, same math, run independently per universe so neither can affect the other.
Vantage isn't telling you to buy or sell. It's flagging a spot where price may be turning — a level worth looking at, nothing more. What you do with that, including which direction if any, is entirely your call.
Both universes run the same three timeframes. This is built for swing positions, not scalping — 15m and 1h were tested on both stocks and forex and dropped from each: too noisy to be worth including. Monthly was tested and dropped on the stock side too: even the oldest, most stable names only have enough real history to cover a fraction of what the detector needs to warm up, so it would only ever produce false confidence, not real signals.
Stocks trade roughly 9:30am–4pm ET, Monday–Friday — Vantage only watches during real trading hours there, nothing fires overnight or on weekends. Forex trades nearly continuously — Sunday 5pm ET through Friday 5pm ET — so that column watches a much wider window, closed only for the Friday-evening-to-Sunday-evening gap. Either way, a closed market never gets mistaken for a quiet one.
Rare by design, same as Fulcrum. Stocks: expect roughly one to three entries a day across the whole universe and all three timeframes combined. Forex is a much smaller universe (7 pairs vs. 518 tickers) and fires less often — expect an entry only every week or two, not every few days. If either list looks quiet, that's the intended behavior, not a bug.
Same two-stage design as Fulcrum. A pivot can't be confirmed the moment it forms — it has to survive a stretch of later trading first, longer on slower timeframes. So Vantage speaks twice: first an early warning, marked EARLY · UNCONFIRMED, sent as soon as a pivot has formed and survived its first test; then a follow-up, either ALL GATES PASSED or DIDN'T HOLD. Both are always sent.
It does not mean the move is guaranteed, and it does not mean price is going up or down 100%. It means every requirement the signal checks was met and the pivot then held afterward. That describes something that already happened, not a forecast of what comes next.
A stock needs a couple hundred closed bars behind it before any of this can be measured. A recently listed one — or one still short on 4H history from this feed specifically — stays quiet until it has enough. That's missing history, not a missing signal.
When it calls a pivot, it's usually right about the general area — but treat that price as a zone, not an exact level to trade off of. Price often keeps going a little past it before actually turning, or the real turn lands a few ticks earlier or later than the number shown. Give it room — same on the stock side as the forex side.
This flags where a turn may be forming. It is not a buy or sell signal, not a prediction, and not financial advice. It reports a condition that has occurred, and says nothing about what price does next. Trade your own decisions.
Tap the 🪲 at the top of the page to report something broken or confusing. Your message goes to the people who run Tephra Terminal. You can add a Discord or email if you want a reply, and before sending you can see exactly which technical details (tab, coin, window size, recent errors) go with it.
A random ID this browser generated for anonymous usage counts — no account, no personal data, and it never leaves this site.
…
@scottytradesSS4 (Twitter)
Join Tephra Terminal Discord — questions, bug reports, or to hear when new features land. Members also get free $CRUM tokens.
Something broken, wrong or confusing? Tell us what happened. Your report goes to the people who run Tephra Terminal.
A tracker of large on-chain money. It watches for coins that pump, follows the wallets that received them in size, and shows large trades and exchange movements across many chains. It reads public blockchain data only, and it is information, not advice.
Every hour ORCA checks for coins up 15% or more over 2 days. Within 6 hours of the pump it looks for wallets that received $250K+ of that coin. Those wallets are followed for 7 days. A single entry of $50M or more is skipped: at that size it is a treasury or token-unlock move, not a trader's purchase.
Skipped on purpose: exchange wallets, DEX pools and routers, bots (dozens of trades a day), and coins that arrived from another private wallet — a wallet-to-wallet move says nothing. Only purchases and exchange movements count.
When a followed wallet sells the pumped coin and buys a different one (over $100K), it appears here with a pop-up, a badge on the ORCA tab and a chime (the same Sound toggle as Fulcrum). "Sent to an exchange" is shown as cashed out. A move through a bridge is followed to the other chain when it lands. This list is often empty — that is normal.
On a phone, Big buys and sells comes first, right under the chain list, followed by Rotations, the exchange flow calendar and watched wallets. A Jump to row (Buys & sells, Rotations, Calendar, Wallets) stays pinned under the header, so every section is one tap away. On a computer the layout is unchanged.
One box per day: the number is how much of the coin left exchanges minus how much went in (plus means more left). The colour compares that day with the previous 30 days — green means more left than usual, red means less. Tap a day to see how unusual it was, how the price moved, and the day before. Use the arrows to change month and Copy as image to share one.
Pick one or several chains. The number beside each is wallets followed plus every move it lists for it ($5K or more on DEX networks, $25K or more on native chains). Only $250K+ moves are listed by default; choose Under $250K to see every smaller move ORCA keeps quietly so every chain builds a record. Greyed chains have nothing yet. Your picks, size filter and Buys/Sells choice are remembered on this device.
Big moves and pump flags on your starred coins also appear in the “Since your last visit” card and the ★ pill on the Heatmap tab (see its i button).
Open any big buy or sell and tap the camera to copy it as an image, or use Copy as image above the list for the latest ten. The calendar has its own Copy as image for the month you are viewing. The status line at the top turns amber if the collector stops updating.
ORCA reports what wallets did on public blockchains. It is not a buy or sell signal, not a prediction, and not financial advice. A big wallet buying, or selling, is not a reason for you to trade. Trade your own decisions.
Tap the 🪲 at the top of the page to report something broken or confusing. Your message goes to the people who run Tephra Terminal. You can add a Discord or email if you want a reply, and before sending you can see exactly which technical details (tab, coin, window size, recent errors) go with it.
A random ID this browser generated for anonymous usage counts — no account, no personal data, and it never leaves this site. Quote it if you report a problem.
…
@scottytradesSS4 (Twitter)
Join Tephra Terminal Discord — questions, bug reports, or to hear when new features land. Members also get free $CRUM tokens.
An on-demand lookup for any ticker, not just Tephra's tracked coins — pulls real candle and volume data live the moment you ask, and runs it through the same Support/Resistance, Value Area, and liquidity-cluster math the Heatmap tab already uses.
The badge at the top (Bullish / Bearish / Neutral) is a plain, rule-based read of market structure — NOT a Fulcrum or Vantage signal, and not investment advice. It's built purely from the 4H, Daily, and Weekly timeframe cards below it: each timeframe is scored on a handful of factors — price vs. that timeframe's value area, support/resistance balance, OBV vs. price (a divergence, such as OBV rising while price falls, counts double), the 50 EMA trend, swing structure (higher highs and higher lows, or lower highs and lower lows), RSI extremes, and how price reacted at the nearest support or resistance level — and the three timeframe scores are added together for the overall call. Every point in every score is spelled out as a plain-English bullet under that timeframe's card — nothing is a black box.
The exact cutoffs: a single timeframe reads Bullish at a score of +2 or higher and Bearish at −2 or lower — anything in between is Neutral. The overall badge uses the three timeframes' scores added together, at +3/−3. If a coin doesn't have enough real history on any of the three timeframes, the badge reads Unavailable rather than guessing.
A one-time snapshot of the order book taken the moment you run the lookup — not a live feed. Flags resting bid and ask orders sized between $500K and $5M (not dust, not a whale's entire position) on whichever venue Vade already trusts most for that coin's price. Because it's a single snapshot, a flagged order can still get pulled or filled moments later — this isn't spoof-detection, just what the book looked like right at that instant. Only shown when a real order in that range actually exists on the book.
A real Open Interest read against price action over roughly the last 24-30 hours, on whichever venue has a real OI history for this coin. This is a descriptive read, not a signal — it's not folded into the Overall bias score above, and it never claims certainty. The plain logic behind it: OI only rises when a genuinely new long and a new short both open, so OI moving WITH price (both up, or both down) tends to mean fresh positioning is backing the move; OI moving AGAINST price (price up while OI falls, or price down while OI falls) tends to mean existing positions are just closing out — short-covering or long liquidation — rather than fresh conviction. Every read spells out the real % change in both series and what that specific combination could suggest, in plain language, the same way the timeframe cards below explain their own reasoning.
Each small pill under a timeframe is one level: a red R is resistance, a green S is support, from real swing highs/lows clustered from that timeframe's own price history. The overall Support/Resistance section below the cards is different: those levels come from the coin's long daily (up to about four years) and weekly price history (not just the current cards' own windows), scored by how sharply price rejected the level afterward rather than by touch count alone — a level tested once with a sharp reversal can outrank one brushed several times with only a mild bounce each time. The ×N next to any chip is still the raw touch count — how many separate swing points clustered into that one level.
The Value Areas table (Day/Week/Swing), Liquidity clusters, and PDH/PDL/PWH/PWL are the same math the Heatmap tab uses for its own fixed windows — only available when this coin has a real futures/perpetual listing somewhere (see Limits below). The overall Support/Resistance list is broader than that: it combines the coin's full daily/weekly key levels with every level already found on the 4H/Daily/Weekly cards above, deduplicated into one list — not capped by count or by how far a level sits from the current price, so it still populates even for a coin with no futures listing anywhere, using the per-timeframe levels alone.
Real recent headlines (last 4 days) that actually mention this ticker, pulled from a set of crypto news sources — each one links straight to the original article. When at least one headline is found, a 1–2 sentence AI summary sits on top, built only from those real headlines. If nothing real is found for this coin, or the summary can't be generated, this section is simply left off the page rather than showing anything invented. Coverage depends entirely on what's actually being written about a coin, so smaller/lower-profile tickers may show nothing here.
The History button shows every past lookup — ticker, bias verdict, price, and a real timestamp — newest first, kept up to the last 300 lookups. The × button next to a result's price clears it and gets the search bar ready for the next ticker.
Tap the ☆ next to a result's price to save the coin. Starred coins show up as one-tap chips under the search box. Your list is stored only in this browser and is private to you.
Tap the 🪲 at the top of the page to report something broken or confusing. Your message goes to the people who run Tephra Terminal. You can add a Discord or email if you want a reply, and before sending you can see exactly which technical details (tab, coin, window size, recent errors) go with it.
Checks futures/perpetual markets first; if a coin has no futures listing anywhere, it falls back to spot-market data instead so a real, actively-traded coin still isn't left out entirely. A coin only reachable via that spot fallback still gets a full three-timeframe bias, RSI, EMA, OBV and Support/Resistance — it just won't have the Value Areas table, Liquidity clusters, or PDH/PWH section, since that data only exists for leveraged markets. Up to 3 lookups can run at once app-wide; a 4th request while those are in progress will ask you to wait a moment.
One shared room — every logged-in Tephra user, all in the same place. There's no DM or separate rooms; anything you send is visible to everyone else signed in.
Type @ and a username to tag them — every registered Tephra user shows up to pick from, not just people who've already posted. A tag is highlighted in the message, and if it's your username, that whole message gets a highlighted border so it's easy to spot in a busy room.
You can send one message every 30 seconds. The Send button disables and counts down while you wait — this keeps the room readable and stops any one person from flooding it.
Messages are capped at 500 characters. The room keeps the most recent 500 messages app-wide; older ones roll off.
5eow3pr9ZCW3Jg9UZMxKY2oQdMNKhEwSLRM7thww3uz5 0xa97aDadb38df0BA8377Aac44Dca4F8c7DE7765E1 It's a map of where leveraged futures positions are likely to get force-closed ("liquidated") if price gets there. Bright horizontal bands mark price levels with a heavy concentration of estimated liquidations — these levels often act as magnets, since a fast move into a dense band can trigger a cascade of forced buying/selling that accelerates price toward it.
Bands build up where estimated liquidations cluster as trading happens. A band clears once price actually trades through it — those positions would already be gone, so there's nothing left to estimate.
Each range is built from its own window of history rather than being a zoom of the same picture, and longer ranges reach further back. So a price zone can show a dense band on a longer range and nothing on a shorter one — not because the data disagrees, but because the activity that built it happened before the shorter window begins. If you don't see a band you expect, try a longer range before assuming it isn't there.
This is not a feed of real liquidation orders — no exchange publishes those directly. It's the same estimation approach used by Coinglass and similar tools: real market data run through the actual liquidation math. Treat brightness as relative likelihood, not a guaranteed dollar amount.
None of this is financial advice. Tephra Terminal shows estimates and observations about market structure. It does not tell you what to buy, sell, or hold, it makes no prediction about where price goes next, and nothing in it should be treated as a recommendation. Trade your own decisions.
Not liquidation estimates — optional toggles kept deliberately separate from the colored heatmap so neither signal blends into it. S/R draws white dashed lines across the chart at real key levels from the coin's daily (up to about four years) and weekly price history, ranked by how sharply price rejected each level afterward rather than by touch count alone, and labeled with the touch count (e.g. "×7"). PDH/L (amber) and PWH/L (deeper amber) mark the prior day's and prior week's high and low as tags on the price axis rather than lines, so they stay legible without covering the bands. Each rolls over on its own within a few minutes of that period closing.
On a phone the lines (S/R, PDH/L, PWH/L and the value areas) and the drawing tools are behind the Tools button, next to the coin and timeframe. Tap it, switch lines on or off, or draw, then close it to see the chart. The number on the button is how many lines are on. The five tabs sit along the bottom of the screen, and the bug-report button lives in this panel under Report a bug. Nothing changes on a computer.
The Footprint button beside Heatmap swaps the chart for an order-flow view. Each candle is split into price rows. On every row the left number (red) is volume sold by market sellers and the right number (green) is volume bought by market buyers, in BTC. Under each candle, Delta is buy volume minus sell volume for the whole candle and Volume is the total traded. The yellow box marks the point of control, the row with the most volume in that candle. A bordered cell is an imbalance: one side is at least 3× the opposite side one row away (buys are compared with the row below, sells with the row above); three or more in a row get a solid bar at the candle edge. Candle sizes run from 1 minute to 1 day.
Data is Bitcoin perpetual trades. Binance is the default, with history back to the end of 2019. Bybit, OKX, Bitget, Hyperliquid and Gate can be added or summed from the exchange picker; their history only starts from when Tephra began collecting it, so a combined view only reaches back to where every picked exchange has data. It shows who was aggressive, not where price goes next, and it is not financial advice.
In the Footprint view the chart fills the window and carries up to two extra panes beneath it. The buttons at the top right choose what sits down there: OI, CVD, BOTH or OFF.
The dots on the chart. These are the order book heatmap — real resting limit orders on Binance's USD-M perpetual book, bucketed into $25 price steps and sampled every 10 seconds. Brighter means more size resting at that price: dark purple is light, yellow is the heaviest. A bright row that price is drifting toward is a wall of resting orders; one that vanishes was pulled or eaten. Order book history cannot be backfilled from any public source, so this record only grows forward from when it started — zoom out past what has been collected and the chart falls back to the modelled Liquidation Heatmap instead. The legend always names which of the two you are looking at, and hovering a dot puts that price bucket and the size resting in it — in dollars — on the same legend line. The OBH button hides them.
Open Interest pane. Open interest drawn as candles on the same time axis as the chart above, so a move in price lines up with the change in positioning beside it.
CVD pane. Aggregated spot cumulative volume delta across Binance, Coinbase, OKX, Bybit, Kraken, Bitfinex, Bitstamp and Gate.io. Each lower-timeframe candle is classified as buying or selling by whether it closed above or below its open, the signed volume is summed across all eight venues, and the running total is plotted. The line is blue while it sits above its own 20-period average and purple below, with the gap between the two filled. It resets at the start of each day. Note this is spot flow — the Delta and Volume numbers under the candles are perpetuals, so the two can disagree, and that disagreement is often the point.
Working with the panes. Each pane has its own vertical scale. Drag the price scale beside a pane to zoom just that pane, drag inside it to move it, and double-click either to re-fit. Drag the divider between panes to resize them. The drawing tools work inside the panes as well as on the price chart, and a line drawn in a pane is anchored to that pane's own values, not to price.
Three toggles in the chart tools — VA/D for the previous day, VA/W for the previous week, and VA/S for the current swing. Each draws three lines across the chart: the price where the most volume traded (POC, solid) plus the top and bottom of the range holding 70% of that volume (VAH / VAL, dashed) — the standard value-area convention. Day and week are calendar periods; VA/S instead uses the range between the dominant confirmed swing high and swing low on the 1h chart over the last 7 days (a volatility-scaled ZigZag, not a fixed lookback), so it tracks the market's current structure rather than a clock. All start switched off, so the chart opens uncluttered. These mark where the market spent most of its time rather than where positions were forced out, which is why they often act as magnets or as the edges of a range. Day and week roll over on their own as each new daily and weekly candle begins; swing updates whenever a bigger pivot confirms or the current one ages out of the 7-day window. They're approximated from candles rather than tick data, so read them as close, not exact.
At the bottom of the chart tools, and not a liquidation estimate — it is ordinary momentum context for whichever coin is selected, so you can see how the same coin is behaving on other timeframes without switching charts.
A dash means there is not enough history yet — the weekly row needs roughly a year of trading before RSI(14) and a 50-week EMA mean anything, so a recently listed coin stays blank there and fills in on its own.
The table refreshes every few minutes and follows the coin picker. It is drawn from the same candle data as everything else, though on the slower timeframes your own charting platform may differ by a fraction depending on which venue it prices from.
By default the heatmap combines every exchange that lists the coin. The Exchanges row in Chart tools lets you look at one exchange on its own (the only button) or add a few together. The list shows each exchange's share of that coin's futures volume, and only exchanges that actually trade the coin appear, so it is shorter for smaller coins. A vol tag means the exchange adds to the volume weighting but has no open-interest history. When you pick exchanges, the band weighting from volume and open interest uses only those exchanges; the price candles, support/resistance levels, leverage model and long/short split stay based on the combined market. A single exchange's view is thinner than the combined one, so treat it as a lens, not a replacement.
The Timeframes row in Chart tools is normally on Single window, the map you pick with the range button. Switch it to Aggregate and the map also marks the biggest liquidation clusters from the 12H, 24H, 3D and 7D windows, each with its estimated liquidation volume (the same figure the hover tooltip shows) and the windows it appears in. Only levels price has not already traded through are shown. Pink lines above price are short liquidations, teal lines below are longs. A level listed in every window has been building at every timescale; one listed only under 7D is older. Tick or untick windows to change what is included. It follows your Exchanges choice and is an estimate like the rest of the map.
Around 170 coins: the top 200 by market cap with stablecoins removed, then filtered by how much futures activity each one actually has. Both filters matter: a coin can be large and still trade too thinly for the estimate to mean much, and below that floor the bands are mostly noise rather than signal. Futures volume is summed across up to ten major venues rather than read off the largest one — plenty of coins do most of their business away from the biggest exchange. Coins with no real futures market are excluded entirely, however large they are: there is nothing there to liquidate.
The list refreshes itself weekly, but it is built to change slowly. A coin joins when it climbs into the top 200, and then stays until it falls out of the top 220. Those two thresholds are deliberately different: with a single cut-off, every coin sitting on the line would drop out and rejoin on a one-place move and the list would reshuffle every week. Having to lose twenty places to leave means a coin you follow stays put unless something real has changed.
Coverage isn't identical across the list. Most coins are priced from eight to ten venues, but a handful trade on only three to five — and that isn't purely a size question: some mid-caps are listed thinly while smaller names appear nearly everywhere. Where a coin has fewer sources its heatmap is built from less data and is correspondingly rougher. Read those as a coarser picture rather than a wrong one.
On the heatmap, a panel on the right (on a phone, a card under the chart with a small map) shows an estimate of where today's open futures positions were opened, on the same price axis as the chart. Longs and shorts use your theme's two accent colors (the header at the top of the panel shows which is which). Solid bars are positions that are underwater right now (longs opened above the current price, shorts opened below it); faded bars are in profit. A big solid stack means a crowd that is hurting and may be forced out if price keeps moving against them.
Exchanges don't publish individual positions, so this is inferred: Tephra replays about 20 days of hourly open-interest and long/short-ratio history from Binance, treats new open interest as opened at that hour's price, and treats shrinking open interest as closed across all positions pro-rata. Positions older than the window are counted but cannot be placed. The long/short split comes from Binance's top-trader positioning applied to all open interest, and exchanges don't say which positions close, so each underwater % is shown with a range across three plausible assumptions (pro-rata, oldest first, newest first). The two sides rarely change order, but the exact numbers can move by 10 points or more.
It is a map of crowding, not a signal. In a check across 20 coins over the last ~20 days it did not predict where price went next, so use it as context for the chart, not as a trade trigger. Not financial advice.
Tap the ☆ next to a coin in the coin picker, or the star beside the picker, to save it. Starred coins appear in a Watchlist row at the top of the picker. It holds up to 20 coins. The app also reopens on the tab, coin and timeframe you last used. All of this is stored only in this browser: nobody else sees your list, and it does not follow you to other devices.
If you've been away for 30 minutes or more, a small card in the top-right corner lists what changed on your starred coins while you were gone (it looks back at most 7 days): Fulcrum signals (new, confirmed or failed, with the result once known) and ORCA wallet moves over $1M and pump flags. Tap a coin to open its chart, or a chip to jump to that tab. If you leave the app open, it checks your starred coins every few minutes and shows a small ★ pill in the top-right when something new happens; tap it to open the same card (it never pops open by itself). With no starred coins it just shows how many new Fulcrum signals appeared. It is a summary of what happened, not a recommendation. The "last visit" time is kept only in this browser and nothing is sent to the server.
With a mouse, double-click anywhere on the chart to screenshot it with that exact spot marked — price, time and estimated volume baked into the picture, so a shared screenshot carries the number you were pointing at instead of an unlabelled band. On a touchscreen there is no hover to read from and a double-tap resets the zoom, so use the camera button instead; it copies the chart as shown. The coin, timeframe, refresh, zoom-reset and camera controls all sit together at the top of the page; the tools list below the chart is only for things drawn on it.
Tap the 🪲 at the top of the page to report something broken or confusing. Your message goes to the people who run Tephra Terminal. You can add a Discord or email if you want a reply, and before sending you can see exactly which technical details (tab, coin, window size, recent errors) go with it.
A random ID this browser generated for anonymous usage counts — no account, no personal data, and it never leaves this site. Quote it if you report a problem, or send it to the site owner to be whitelisted from rate limits.
…
@scottytradesSS4 (Twitter)
Join Tephra Terminal Discord — questions, bug reports, or to hear when new features land. Members also get free $CRUM tokens.
Tephra Terminal runs on a home server, not a funded infrastructure budget. If you find it useful, contributions toward hosting and continued development are appreciated.
USDT/USDC — Solana
5eow3pr9ZCW3Jg9UZMxKY2oQdMNKhEwSLRM7thww3uz5
USDT/USDC — Ethereum, Base, Polygon, Avalanche, Arbitrum
0xa97aDadb38df0BA8377Aac44Dca4F8c7DE7765E1
Clipboard copy isn't supported on this device/browser. Tap the button below to download instead, or press and hold (or right-click) the image and choose Save Image / Copy Image — then paste it into a message to share it.