Tokenized equities trade in dozens of pools that disagree, and the one advertising the best price is often the one that cannot fill you. Hush reads every venue on the chain and quotes all of them at the size you are actually trading.
AAPL3 venues
AMD1 venue
AMZN2 venues
GLD4 venues
GOOGL3 venues
INTC1 venue
META2 venues
MSFT3 venues
MSTR2 venues
NFLX1 venue
NVDA4 venues
PLTR2 venues
QQQ3 venues
SLV2 venues
SPY4 venues
TSLA3 venuesThis quotes, it does not trade. Every venue is read at the block shown and walked at your actual size, so the number above is what that pool would pay, not a spot price.
The problem
Here is QQQ against USDG, sorted by price the way a book is, beside the depth behind each price and what $10,000 actually fills there.
| Venue | Advertised price | Depth to 1% | $10,000 fills |
|---|---|---|---|
| v4 · 0.048%not routable | $715.1225 | $7 | 0.0112 sh |
| v3 · 0.050% | $716.4981 | $689,686 | 13.956 sh |
| v4 · 0.350% | $718.8629 | $11,910 | 13.867 sh |
| v3 · 0.300% | $719.6891 | $4,958 | 13.750 sh |
The cheapest row fills 0.0112 shares on $7 of depth. The row that fills best gives you 13.956, for the same asset at the same block. Sorting by price is how you end up in the first one.
Block 58,962,250, 1h ago. A snapshot. The panel above is not, and re-reads every pool as you type.
What it costs
1,248× apart. v4 · 0.048% advertises 19 basis points cheaper than v3 · 0.050%, and works out at $894,170 a share against $716.55, because there is nothing behind the price. Both come from walking those two pools’ own liquidity at block 58,962,250.
Pools initialised at an arbitrary price that never traded still answer a quote. One offers NVDA at $128,373,698 a share on a dollar of liquidity. Assets are pinned by address, and 4 pools that were not pricing their asset at all were dropped.
Not spot. Hush reads each pool’s own state and walks its liquidity tick by tick at the size you are trading, the way the pool would when it fills you. Against the same walk running on-chain, 225 of 265 quotes agree to the wei.
One venue, or several when splitting genuinely wins. At retail size it usually does not, because the cheapest venue is normally the deepest, so the solver names one place rather than carving the order up for show.
Every venue read is published beside the one that won, with the depth behind each price. The pools too thin to send an order to stay in the table, because hiding them would hide what you are being warned about.
Across the chain
Every indexed asset quoted on more than one venue, widest disagreement first, from GLD at 196 basis points down to PLTR at 8.
| Asset | Spread · bps | Range | Venues | Deepest venue |
|---|---|---|---|---|
| GLD | 196 | $402.61 → $410.49 | 7 | $572,598 |
| META | 186 | $649.15 → $661.25 | 5 | $55,937 |
| AMZN | 173 | $252.87 → $257.25 | 4 | $178,420 |
| TSLA | 172 | $367.28 → $373.59 | 5 | $107,800 |
| SLV | 151 | $60.83 → $61.75 | 4 | $58,678 |
| NFLX | 132 | $76.38 → $77.39 | 2 | $21,505 |
| AAPL | 123 | $317.48 → $321.39 | 7 | $152,979 |
| MSFT | 122 | $492.86 → $498.88 | 5 | $121,719 |
| AMD | 102 | $521.84 → $527.16 | 3 | $14,817 |
| SPCX | 101 | $147.28 → $148.78 | 5 | $207,742 |
| MSTR | 100 | $133.77 → $135.10 | 5 | $42,518 |
| NVDA | 66 | $223.89 → $225.37 | 7 | $874,573 |
| QQQ | 64 | $715.12 → $719.69 | 4 | $689,686 |
| SPY | 50 | $763.52 → $767.36 | 4 | $4,686,583 |
| GOOGL | 43 | $331.31 → $332.72 | 4 | $62,004 |
| PLTR | 8 | $170.02 → $170.15 | 2 | $14,839 |
FAQ
A displayed price is the price for an infinitely small trade. The moment you buy you move it, and how far depends on how much money is in the pool, which the price does not tell you. In the book above, the worst row is 0.0112 shares on $7 of depth, and it is not even the dearest.
That every venue on this chain for these assets has been found, read from its own state, and quoted at the size you typed, and that the one named is the best of them. Nothing about execution is claimed, because nothing here executes.
The quoting is checked against the same walk running on-chain: 225 of 265 quotes matched to the wei. Of the rest, 9 are orders no pool could fill anyway. Past the end of a pool’s liquidity the two walks stop in slightly different places. On every order that does fill, this engine is within 3.1e-15 of the on-chain result. npx tsx scripts/audit.mts re-runs it.
No, and there is nothing here that could. This site has no contract of its own and no wallet connection. It reads public chain state and does arithmetic on it. You take the answer and trade wherever you like.
Only when it pays for itself. The smallest order where splitting beat a single venue was $1,000, and on 8 assets no size tested made it worth the extra gas, because the cheapest venue there is also the deepest. Most of the value is picking the right venue, not carving the order up, and the panel says which it did.
A quote is a reading of one block, and anyone can move the pool between it and your trade. The deeper the venue, the less that matters. The book covers what the indexer found: the canonical v3 factory and every v4 pool the PoolManager announced. A venue outside that set is not considered, so you are trusting that curation. The script that produced it ships with the site.
