Direct APK Distribution
Quantlow for Android
Your signals, live pair watches, and performance tracking directly on your phone. Every engine fix ships here first.
Version 1.0.25build 27
October 5, 2026Android 8.0 and higher. No Google Play account required.
What Changed in Build 27
- The Account screen no longer prints our bank account number to everyone who opens it. It was written into a sentence shown to every user whether or not they were buying anything, which is payment detail put in front of someone who has not said they are paying.
- Choosing a plan and choosing how to pay are two separate steps now, and both live on the plans page. That is where the amount can be shown for the plan you actually picked, and where your payment reference belongs. The app tells you the transfer route exists and sends you there, and carries no account number at all.
- Everything in 1.0.24 is included. If you are coming from 1.0.23, see those notes as well.
SHA-256 Checksum Verification
256183d69a679f58ccbd0ae6cbddef252efbfbc15d426b54f545fdb1e261ec15
How to Install on Android
- 1Tap the Download APK button above directly on your Android phone.
- 2Open the downloaded file in your browser downloads or file manager. If Android asks whether to allow installs from your browser, tap Allow.
- 3Tap Install. If updating an existing version, your login session and watch preferences are preserved.
- 4Open Quantlow and sign in. You can then enable biometric authentication.
Previous Builds
View full changelog1.0.24build 262026-10-05
- Alerts now update themselves. The Watches screen refreshes about once a second while you are looking at it, so the countdown, the live price and the drift move on their own and you are not pulling to refresh to find out whether anything changed. That was the actual complaint and it was a fair one.
- What the word “live” can honestly mean here, because it is worth being exact. Two separate things were slow. The screen not updating itself was ours, it was free to fix, and it is fixed. How fresh the PRICE is belongs to the data provider, and a closed one-minute candle reaches us about twenty seconds after that minute ends. We measured it: at every check up to sixteen seconds past the minute the bar was still missing, and it was reliably there by twenty-three. So the price can be up to roughly that old however often we ask, and asking every second would have cost sixty provider credits a minute for a single pair against a ceiling of fifty-five for everything the app does, to redraw the same number twenty times. We ask every five seconds, and only for pairs somebody currently has open.
- The drift line under an alert is an observation and nothing else. It says how far price has moved since the alert, in pips, in the direction the alert called. It does not change the alert, re-order anything, hide anything or predict anything, and a drift going the right way is not evidence the read was right. It is there because you could already see it by watching your broker, and having to switch apps to see it was the only reason it was not on the card.
- New: “I did not take this”, for a trade you logged by mistake. Tapping “I took this” on the wrong card used to be permanent, which meant the honest options were to leave a trade you never placed in your record or to ask us to delete it.
- It has one hard limit, and the limit is the point. A trade that has already settled cannot be removed. That is enforced by the database itself rather than by the app, so no future version of the app, and no mistake by us, can get around it. If a settled trade could be removed, every record on every card would become a number you can edit once you know how it turned out, and then none of them would mean anything. Before it settles, a mis-tap is just a mis-tap and you can take it back.
- A removal is recorded as a removal with the reason, not erased. Your counts and rates are worked out from what you actually took, and the removed row stays visible as history.
- Session history. Every session you have finished, with what you planned, what you did, and the trades in it. Previously a session ended and took everything in it with it.
- The session setup screen is readable now. It was five cards of explanation with the Start button underneath all of it. Every paragraph was true and worth saying, and together they meant reading for a minute before starting something you already understood. The paragraphs are not gone, they are under “What these do”, word for word. One thing is deliberately still on the screen rather than inside the disclosure: Quantlow cannot stop a trade. Your broker already has it by the time you tell us. The toggle is called “warn me firmly” because that is all it can do, and anyone who never opens the disclosure still needs to know that.
- The correction promised in the last build’s notes is done. The Performance page says “Break-even at your 92% payout” and names the rate it used, worked out from the median of the payouts you actually logged. We had changed it to “the assumed payout” in 1.0.23, which was wrong of us: that figure was always yours, and it is left blank rather than guessed when you have logged none.
- The money editor in the floating bubble takes a stake and a payout where you are already looking, instead of sending you to another screen to record what you just did.
- Plans and upgrading now work from inside the app, and the limits shown come from the server rather than being typed into the app, so the page cannot drift out of date from what your account actually allows. What a paid plan buys is access, limits and tools. It does not buy better reads, and no plan is unlimited, because every scan costs real money at the vision provider.
1.0.23build 252026-10-02
- Watch alerts now carry the same measured record the signal cards carry: what reads like this one have actually done, with the sample size. A watch alert had no record at all before, on the grounds that it had none of its own and borrowing the signal feed’s would be an overclaim. That was right about the borrowing and wrong about the absence. A watch alert is the same engine’s read, it settles into the same tables, and it was already being scored. The line was simply never shown to you.
- Fixed, and this one was telling you something untrue. The binary alert card read “Needs 56% of these to win, at your payout, just to break even.” That number was never your payout. It came from the configured default, so the card was showing an assumption back to you as your own figure. It now says “at your 92% payout” only when a live session actually recorded one, and “at the standard 80% payout” otherwise.
- Gold is measured as gold. Until now every non-binary record pooled gold in with the currency pairs, so a gold alert printed a number earned by EUR/USD and friends and labelled it gold. Gold has 14 settled reads of its own, below the minimum sample, so gold cards now say “New setup, building its record.” That is less satisfying and it is the truth: gold’s own 14 reads won 4, well below the 40% the borrowed figure was showing you.
- New: a heads-up when something is scheduled. Alerts say when a high-impact release is due within the next half hour for either currency of the pair, for example “US 3 releases in 20 minutes”. Where several land in the same minute it counts them rather than naming one, because the calendar does not rank them and picking one would imply it was the one that mattered. Gold uses US events, since it is priced in dollars.
- Alerts also say when either currency has a bank holiday, because the market is thinner than usual. A daylight-saving clock change is filed under the same heading by the calendar we read and is deliberately NOT reported as a thin market, because it is not one.
- Both of those are information and nothing else. They do not hide an alert, hold one back, re-order anything or change a single number. Reads near the news look obviously worse, and so did reads on a barely-moving market, which we measured: those tie more often rather than lose more often, so filtering them would have cost money. Every read now records whether it fell near an event, so the question gets answered with evidence instead of instinct.
- Correction to this build: we also changed the Performance page from “Break-even at your payout” to “at the assumed payout”, and that was a mistake on our part. That figure was always yours: it is worked out from the median of the payouts you actually logged, and it is left blank rather than guessed when you have logged none. The next build restores it and names the rate, so it reads “Break-even at your 92% payout: 52.1%” and you can check it.
- In the floating bubble, the scheduled-event line appears on every alert, because “US 3 releases in 5 minutes” while you are about to place a one-minute trade is exactly what the bubble is for. The measured record appears only when there is a single alert: it is two lines, and on a stack of three it would push the pair and the clock down the card, and those two have to stay the loudest things there.
1.0.22build 242026-10-02
- The Session warnings now reach you at the moment you log a trade, wherever you log it: a watch alert, a forex alert, the performance screen or the home-screen widget. Before this they only appeared if you opened the Session screen yourself, which is the one place you are not looking while you are trading.
- They warn and then get out of the way. Quantlow never places a trade, so everything you tap in this app is a RECORD of something your broker already has, and blocking it would not stop a trade, it would only stop the trade being measured. The earlier wording suggested the app could hold a trade back. It cannot, and it no longer says it can.
- If you asked to have your alerts held once you hit your planned number of trades, the Session screen now says so at the top: “Alerts paused until you end this session”. It also says the two things that matter beyond the fact itself, that everything is still being recorded and scored as normal so nothing is lost from your record, and exactly how to lift it. A pause you cannot find or lift is not a tool.
1.0.21build 232026-10-01
- The confidence percentage is gone, from every screen, the home-screen widget and the alert notifications. It is not hidden or demoted, it is removed. That number was the engine rating its own read, and we measured the rating against settled outcomes: across 1,174 settled binary trades the four setup groups all sit within a few points of a coin flip. So a card reading “69%, SOLID” was promising an edge the record does not support. Worse, the group the engine rated highest was not the best one, so the number was not merely uninformative, it pointed the wrong way.
- In its place, what setups like that one have actually done: “Setups like this won 7 of 10 (37 trades)”. The sample size is always there, because a rate without one is not a measurement. Below the minimum sample it says “New setup, building its record” rather than inventing a figure.
- Underneath it, the bar that record has to clear: “needs 6 of 10 to break even”. Without it a low number is ambiguous. Three of ten is losing at 1.4 to 1 and perfectly fine at 3 to 1, and nobody should have to work that out from the card.
- On binary the record is one number for the whole market, and the card says so. The four setup types have not been measured to differ, so showing four separate labels over one number would be a distinction the data does not support. On forex the groups genuinely do differ, so those are shown separately.
- The record is deliberately not colour-coded by how good it is. A green seven-of-ten beside a red three-of-ten would put the app back in the business of grading a setup for you, which is the habit the confidence number taught.
- New: Session, under Account. Set how many trades you plan, your broker payout and your usual stake before you start, and the app tracks the session against them.
- The trade count is the first thing it asks and it starts empty, labelled “No limit”. We replayed the whole broker history through every stop rule we could think of, and the only thing that reliably helped was taking fewer trades, because the average trade loses and so volume is the cost. That is arithmetic about how many you take, not a view on which ones, and the limit is yours to set rather than ours to suggest.
- Your payout now changes what the cards say. At 80% you need 6 of 10 to break even, at 92% you need 5 of 10, and the same record of 5 of 10 is losing under one and level under the other. Until you enter it, cards say “the standard 80% payout” rather than treating 80% as yours, because an assumption should never be shown back to you as your own figure.
- It will tell you when a stake is 1.5 times the size you planned. A bigger stake does not change how often a setup wins, it changes what the next loss costs.
- It will tell you when you have had three losses in a row, and then tell you that this says nothing about the next trade. We checked your own history for the streak everyone believes in: 50.0% of trades after a loss lost, against 50.0% after a win. The losses do not cluster. So the app shows you the run and refuses to read anything into it, and a run of losses never locks you out, whatever else you have switched on.
- Locking yourself out is off by default and only you can turn it on. Everything else warns and then gets out of the way.
- If you are ahead, it says so and offers stopping as a choice, and says plainly that we have no measurement showing that stopping while ahead earns more than carrying on. It is advice, not a saving we are claiming.
- Where a logged trade has no amount against it, the session result says so rather than showing a partial figure as the day's total.
- Fixed, in a sense: the free plan advertised “Dual-engine signals”. The two AI engines have been unavailable since early August and every signal since then has come from the rules engine. The scan card also described the rules engine as a stand-in used “while the AI engines are unavailable”. It is the engine, and it now says so.
- Watch alerts no longer show a confidence figure. They have no settled record of their own yet, and borrowing the signal feed's would be a different kind of overclaim.
1.0.20build 222026-09-28
- Fixed: recording a forex result never worked, for anyone. Two separate faults made the same error, “a trade cannot be closed before it was entered”. The first: when you came back to say how a trade ended, the app sent its own clock for the closing time, and the server compared that against the current moment instead of against the entry time it already had. A phone's timestamp is always a fraction of a second behind by the time the request lands, so every close was refused. The second: the entry time was stamped the moment you tapped “I placed this”, which is right for a one-minute trade and wrong for one that runs for hours.
- You now enter the times the trade actually opened and closed, in your own clock with the time zone named, so you can copy them straight off your broker's statement. Leave the closing time as it is and the server stamps it, which is the honest default for “it just closed”.
- After you record a result it tells you what went in, including the R worked out from your prices. That is the number that catches a price typed into the wrong box, and one was: an entry price in the exit field would have recorded a flat trade without a murmur.
- A settled trade now shows its result properly: the outcome in words, both prices, both times and the R. It used to show the raw word the database stores and no exit price at all, so there was no way to check on the card that what you typed was what you meant.
- The four outcome buttons fit. “Break-even” was being squeezed into a single letter per line down the side of the screen.
- Old forex alerts move out of the way. An H1 plan stops being actionable when its hour is out, and the screen went on offering “I placed this” on alerts from five hours earlier. Live ones are at the top; the rest fold into a list you can open. Anything you have taken and not closed stays at the top, because an open trade still needs closing.
- “No measured record yet” appears once, above the list, instead of on every card. It is one number for the whole forex test, and repeating it made a single fact look like a finding about each alert.
- The experimental banner is gone from the forex screen. Nothing about the claim has changed: the record line still says in plain words when there is no measured edge.
- Messages clear themselves after a few seconds rather than waiting to be dismissed, and on the forex screen they now appear at the bottom where you are looking. Before, a confirmation went to the top of a long scrolling list, so submitting from a card further down looked like nothing had happened at all.
- Fixed: the bubble never let go of a forex alert. Only binary rows checked whether the trade had been logged, so a forex trade recorded on the Watches screen sat in the bubble for the rest of its hour still offering itself. Expired plans are also kept out now rather than appearing for a moment on every restart.
- Fixed: the bubble claimed “25th read on this pair”. The count was adding to itself on every refresh, so it was measuring how often the app checked rather than how often the market did.
- New: a forex home-screen widget. Pair, direction, entry, stop, target and how long the plan stays valid, with no countdown anywhere on it, because an hourly setup wearing a one-minute face is a trade placed on the wrong horizon. Tapping it opens Watches already on forex.
- Gold can be watched on forex now, on the same experimental terms as everything else. It was kept out for being the worst instrument on our record; that turned out to be a measurement error on our side. Gold prices and its risk are shown in dollars, not pips — a live gold alert once read “risk 222721.8 pips”, which was dollars divided by the wrong unit. A gold watch may go several days without an alert; that is the instrument, not a fault.
- The pair is the biggest thing on a bubble row and on a forex card, larger than the direction. Similar tickers sit next to each other on a broker's list and are easy to mis-tap.
