Best Liquidity Sweep Indicators on TradingView (2026), Code-Read.

Short answer: the popular open-source liquidity sweep scripts disagree on three things: which level a sweep takes, whether price must close back inside, and whether they wait for the candle to close. The most boosted, Liquidity Sweeps by LuxAlgo, marks the textbook sweep, a wick through a swing and a close back on the same candle, on the live candle and with no alerts. For a sweep decided on a closed candle that you can alert on, Liquidity Sweep Hunter Algo by AlgoAlpha is the closest in our read, with a stricter rule of its own.

Most lists name liquidity sweep indicators.

This one reads them.

What we checked

A liquidity sweep is a candle that trades through a level where stop orders are likely to sit, a swing high or low or a session's high or low, and closes back on the other side. A raid that keeps closing beyond the level is a break, not a sweep.

We searched TradingView's scripts for ten terms on 26 October 2026, among them "liquidity sweep", "liquidity sweeps", "stop hunt" and "session liquidity", which returned 307 scripts. We took the nine most-boosted open-source scripts with "sweep" or "liquidity" in the name, down to 5,704 boosts, none skipped, and read every line that decides a sweep. Four questions:

• The level: what does a sweep take: a pivot, a fractal, a rolling high or low, a session's high or low?
• The close back: must price close back inside, and how soon?
• When it decides: on a closed candle, or on the live one?
• Alerts: can you set an alert for the sweep itself?

The table

Boosts as TradingView's search showed them on 26 October 2026. ATR(14) is the average true range of the last 14 candles.

Two more of the nine aren't in the table. TradingFinder's ICT Concept, Order Block / FVG / Liquidity Sweeps (9,235) finds its liquidity in an imported library we didn't read, and its own signals fire when price touches an order block. Session Liquidity by Infinity_Trading_ (5,704) draws the highs and lows of two of its three session windows by default, on New York time, but has no sweep test and no alerts. Two scripts with "stop hunt" in the name, Inducement / Stop Hunt by TFO (6,131) and Stop Hunt by Nephew_Sam_ (5,370), are protected (code hidden), so we couldn't read them.

What the code says

LuxAlgo's Liquidity Sweeps is the textbook rule, on the live candle. It finds pivots with five candles on each side, so a level exists five candles after its swing, and marks a sweep when a candle's wick gets through and the same candle closes back. A close beyond ends the level in its default mode; its other two modes keep a broken level and mark a later wick back through it as an "outbreak and retest" sweep. Nothing waits for the close, so a sweep can print mid-candle and be gone, and there are no alerts. It is shared under CC BY-NC-SA 4.0.

LuxAlgo's Pure Price Action version uses swings of swings. Instead of a length, it nests three-candle fractals: at its default "Long Term" depth, a level is a swing of a swing of a swing, confirmed only when the smaller swing after it is. The sweep is the same same-candle wick and close back, on the live candle, with no alerts.

wut_trader's counts a touch, and names three patterns. The code's own title is "SMC Target Liquidity V.35 (Manual Distance Control)". On pivots of 10 candles each side, it marks "X" when a wick reaches the level, a touch counts, and the candle closes back; "SFP" when a close beyond is followed by a close back; and "MSS" when two closes beyond say the level broke. Session times only colour its lines. No alerts, no closed-candle check and no licence line.

Mirage is a signal system built on the sweep. It takes pivots of 21 candles each side, sweeps only levels at most 80 candles old, and scores each sweep 0 to 100 on wick length, how far the close came back, where it closed in the candle, volume and a 4-hour trend; below 50, nothing. By default it then waits up to 13 candles for a close beyond the latest swing with 8 candles each side before it prints Long or Short, with a stop 0.25 ATR past the sweep's wick and targets at 1R, 2R and 3R, one trade at a time. Everything that signals waits for the close, and its 4-hour trend is read one candle back, which keeps it from repainting. Its alerts are for entries, stops and targets, not for the sweep. Its on-chart win rate counts a trade as a win once the first target is touched. It carries no licence line.

AlgoAlpha's Liquidity Sweep Filter doesn't ask for a close back. Any wick through a three-candle fractal that sits on the right side of its trend band counts, weighted by the fractal candle's volume; above 50 on its volume scale, the mark is "major". It decides on the live candle and has six alerts, among them a bullish and a bearish sweep. Without volume data it marks nothing. It is shared under MPL 2.0.

AlgoAlpha's Liquidity Sweep Hunter Algo is the strictest. It keeps rolling highs and lows over 14, 42 and 120 candles, and on a closed candle needs the wick to reach at least two separate levels whose volume-based strength averages 33% or more, then a close back past the nearest one within two candles. It draws an entry at the next candle's open, a stop at the four-candle extreme and three targets up to 3R, and alerts on the sweep reversal. Without volume data no sweep passes its default filter. It is shared under MPL 2.0.

AlgoAlpha's Nadaraya-Watson version means something else by "swept". Its one level comes from a regression oscillator turning, and "swept" is a close beyond it: no wick test, no close back. It decides on the live candle and has ten alerts, two of them for the level swept.

None of the nine has anything gold-specific in the code we read, and none of the seven whose sweep test we read sweeps a session's high or low.

For scripts that draw the killzones' highs and lows themselves, we read the 15 most-boosted killzone and session scripts (best killzone indicators, code-read); none of them tests for a close back inside.

The scripts built on yesterday's high and low are another family; we read the 15 most boosted, and none of them tells a sweep of that level from a break: best previous day high low indicators, code-read.

Ours

Gold Sweep Desk isn't on TradingView's search: its full Pine v6 source comes in Pine Pro Library, $49 once. It is built for gold and for session liquidity: its levels are yesterday's high and low, the Asia range's (20:00 to 24:00 New York) and the London range's (02:00 to 05:00). A sweep counts only in the London (02:00 to 05:00) or New York (07:00 to 11:00) window, when a raid gets at least 0.05 × ATR(14) through a fresh level and a close comes back inside on the raid candle or within the next three; a raid that keeps closing beyond is held. Then it waits for the flip, a close through the nearest opposing fair value gap, before it draws a plan. Everything is decided on closed candles, every clock is New York's, and it alerts on the sweep, the entry at the flip, 1R and the exit. A strategy twin runs the same rule in the Strategy Tester.

The rule, step by step and with a clock you can try, is in the gold liquidity sweep strategy. We tested it twice on 2026 gold data and no setting became a tested default: the one candidate that passed its first period then lost 11.63R over 15 trades in its check period (every row). Looking for a gold indicator rather than a sweep one? The best gold indicators for TradingView, code-read.

Check one yourself

• The level: find the ta.pivothigh/ta.pivotlow call (its two numbers are the candles each side), the fractal test or the session window. That's what gets swept.
• The close back: a sweep test reads like high > level and close < level. With only high > level, any wick counts; with close > level, it's a break.
• The close: look for barstate.isconfirmed around the sweep. Without it, a sweep you see live may not be there when the candle closes.

None of the scripts here, ours included, runs on MT5 or MT4: Pine Script runs only on TradingView.

Limits

One code read, on 26 October 2026, of the versions TradingView served that day. Scripts change, and boosts measure popularity, not quality. We build and sell Gold Sweep Desk, and we have no relationship with any author named here. Nothing here is financial advice.