Densities Scanner

A density is a large resting limit order: a wall in the order book. Bids below the price defend it as support; asks above cap it as resistance. Osiris maintains a live local copy of the order book for every perpetual pair and scans all of them every two seconds, so the table you see is a real-time census of where size is sitting across the entire market.

Every exchange has its own scanner page: use the venue switcher at the top of the densities page (or open /densities/bybit, /densities/okx, /densities/bitget directly) to read any venue's books. The bare /densities URL opens on the exchange your screener is pointed at. Venue pages are open on every plan; no account is needed to read the walls.

Why walls matter for scalping

Price interacts with large resting liquidity in repeatable ways. A big bid wall just under the price gives aggressive sellers something to sell into. If it holds, it becomes visible support; if it is pulled or eaten, the price often accelerates through the level. Walls also act as magnets: market makers and algos route around them, and price frequently drifts toward large liquidity before reacting to it. Knowing where the walls are turns the order book from an opaque stream into a map.

What counts as a wall

A price level qualifies when it is large both in absolute and in relative terms: at least $200K of resting orders, and at least 20× the median order-book level of that pair within 10% of the price. The relative test matters: $300K is furniture in the BTC book but a fortress on a small alt. Up to four walls per side are tracked per pair, so the table stays focused on the levels that dominate each book.

Reading the table

• Side: BID (below price, support) or ASK (above price, resistance).
• Wall price / Distance: where it sits and how far from the current price. Walls under 1% away are the ones scalps interact with; walls 5% away are context, not levels.
• Size: the wall's notional in USD, with a bar relative to the biggest wall in view.
• Strength: how many times larger the wall is than the pair's typical book level. A 100× wall dominates its book even if its dollar size looks modest; strength is the fairest way to compare walls across pairs of very different liquidity.
• Lifetime: how long the wall has existed. This is the honesty metric: walls that sit for hours are usually real interest, while walls that blink in and out are often spoofing. Osiris tracks walls across size changes (walls get partially eaten and refilled) and short disappearances, so a spoofer pulling and re-placing an order does not reset the clock.

Filters

Everything is filterable: side, minimum size, maximum distance, minimum lifetime, minimum strength, plus a pair search. Filters are remembered in your browser: they are a device preference, so they work signed in or not, and each device can keep its own view. A practical starting point for intraday scalping: distance under 1%, lifetime over 15m. That view shows only committed walls close enough to matter right now.

Trading around walls

• Support/resistance plays: a long-lived bid wall just below price is a reasonable invalidation anchor: enter above it, stop below it. If the wall gets eaten and price trades through, the defense failed and the trade idea is dead; exit without debate.
• Absorption: watch the size column live. A wall being consumed while holding price (size shrinking, price stable) is absorbing aggressive sellers; if it finishes the absorption intact, the sellers are exhausted, which is often bullish.
• The pull: a large wall that vanishes without being traded through tells you the player changed their mind, and the level it was defending is now undefended. Fast moves often follow.
• Breakout confirmation: a breakout that eats a major ask wall has real money behind it; a breakout into an untouched wall is running into a ceiling.

Honest caveats

Never treat a wall as a guarantee. Large players cancel orders; icebergs hide true size; and a wall can be bait. The lifetime column tells you how committed the wall has been so far, not what it will do next. Use walls as context and invalidation levels, never as the whole trade thesis.