Confluence Entry Signals on Buildix: Every Closed 5m and 15m Bar, Graded and Pushed to Telegram

Most orderflow signals stop existing the moment you close the tab. Buildix computes confluence entry signals server side, every minute, on the twelve most liquid Hyperliquid perpetuals, across both the 5 minute and 15 minute timeframes. The chart does not need to be open. Your laptop does not need to be on.

That distinction sounds pedantic until the setup you spent a week waiting for prints at 04:00 while you are asleep. A signal that only computes when a browser tab is rendering is a chart indicator with extra steps. A signal that runs on a schedule, grades itself, and pushes to your phone is something you can actually build a routine around.

Why Do Browser Side Signals Miss the Setups That Matter?

Almost every retail orderflow tool computes in the client. The candles arrive over a websocket, the indicator runs in JavaScript, and the alert fires from the same tab. Close the tab and the whole pipeline stops.

That model fails in exactly the conditions worth trading. Thin liquidity hours, weekend sessions, the 3 a.m. sweep that resets funding. Those are the moments where orderflow dislocations are cleanest, and they are also the moments nobody is watching a screen.

The Buildix entry signal engine runs on a one minute schedule on the server. Every cycle it pulls the current bar state for the top twelve Hyperliquid pairs by 24 hour volume, evaluates the confluence logic on closed bars, and writes the result whether anyone is looking or not.

Ranking by volume rather than by a fixed watchlist matters too. The twelve pairs are recomputed from live volume, so when a new market starts absorbing real flow it enters the coverage set on its own.

What Does the Engine Check Before It Calls an Entry?

A signal only fires when several independent orderflow conditions agree on the same closed bar. Any single condition, taken alone, produces the kind of noise that makes people turn alerts off after two days.

Per bar delta is computed from actual trade data wherever the trade window covers the bar, rather than inferred from the candle shape. The difference shows up precisely on the bars that matter: a long lower wick with genuine buy aggression and a long lower wick with passive absorption look identical on price and completely different in delta.

Every signal carries a grade of A, B, or C, plus veto conditions. A veto is not a weaker signal, it is a discard: when the confluence checks point one way and the surrounding context contradicts them, the setup is thrown out instead of being downgraded and shipped anyway.

The engine only evaluates closed bars. The partial bar in progress never produces a signal, which is the reason nothing on the signal history repaints. An alert that appears and then quietly disappears twenty seconds later is worse than no alert.

Why Add a 15 Minute Timeframe Next to 5 Minutes?

Five minute bars are the right resolution for scalping and the wrong resolution for almost everything else. Traders holding positions for hours were getting pinged on moves that resolved before they could act.

The 15 minute layer is built by aggregating the same 5 minute bars the engine already has, aligned to proper 15 minute boundaries. Three complete 5 minute bars make one 15 minute bar. If any of the three is missing, gapped, or still open, the bucket is discarded rather than approximated.

The practical consequence is that the higher timeframe costs nothing in latency or exchange load, because no additional market data request is made. Same data, two resolutions, same engine.

Signal identity is namespaced per timeframe, so a 15 minute signal and a 5 minute signal landing on the same timestamp stay distinct. That sounds like an implementation detail, and it is, but it is the detail that keeps one timeframe from silently suppressing notifications on the other.

How Do the Alerts Actually Reach You?

Entry signal alerts are configured on buildix.trade/dashboard/alerts. Pick the symbols you care about, set a minimum grade, and choose whether you want everything or only the A setups.

Delivery goes to Telegram if your account is linked, and falls back to email otherwise. Each signal has a deterministic identity tied to its bar, so the same setup is never delivered twice, even if the engine reevaluates the same bar on a later cycle.

The alerts sit on the Pro and Whale tiers. The signals themselves are visible in the interface for the tiers that include them, but the push delivery is where the feature earns its keep, because the entire point is not having to look.

One deliberate choice: only the 5 minute signals dispatch notifications right now. The 15 minute layer is available to read, and traders who want it in their notification stream can say so, because turning it on without asking would double the volume of pings for everyone.

How Should You Use a Graded Signal Without Outsourcing Your Judgement?

A grade A signal is a pointer, not an instruction. It says several orderflow conditions aligned on one closed bar in one market. It says nothing about your account size, your existing exposure, or whether a macro print lands in ninety seconds.

The workflow that works: alert arrives, open the pair deep view, check the surrounding context before doing anything. Funding, open interest direction, whether the level in question sits under a dense liquidation cluster, and whether the CVD picture on the higher timeframe agrees.

Grade C signals are most useful as a heat map rather than as trades. A market printing repeated Cs in the same direction across an hour is telling you where flow is building, even when no individual bar is worth acting on.

If you want to see the raw inputs behind the grade, they are all on the pair view at buildix.trade/pair/BTC: CVD, order book imbalance, VPIN, volume profile, and whale activity in one place. The signal points, the deep view confirms.

FAQ

Which pairs are covered? The top twelve Hyperliquid perpetuals by 24 hour volume, recalculated continuously. Coverage follows liquidity rather than a hardcoded list.

Do signals repaint? No. Only closed bars are evaluated, and each signal is keyed to its bar, so a signal that appeared cannot change or vanish later.

What is the difference between a veto and a low grade? A low grade means weak confluence. A veto means the context actively contradicts the setup, and the signal is discarded rather than published at a lower grade.

How fast do alerts arrive after a bar closes? The engine cycles every minute, so dispatch lands within roughly a minute of the bar closing, subject to delivery time on Telegram or email.

Systematic traders do not lose to a lack of ideas. They lose to inconsistency, to the setups that fired while they were doing something else, and to indicators that quietly rewrote themselves after the fact. Fixing that is less exciting than a new indicator and considerably more valuable.

This article is educational and not financial advice. Signals are analytical outputs, not trade recommendations. Perpetual futures carry a high risk of loss, including total loss of margin.