Catalog vs. Quote Engine: 90 Long-Tail Pairs, Measured
We tried 90 long-tail pairs against our own quote engine: zero returned 'pair not supported'. The real wall is the minimum — about $10-13 of value, priced in units you never counted.
The catalog page says 476 assets across 67 networks — 381 distinct tickers, 57 flagged as stablecoins, six requiring an extra ID, ten featured. Catalog size is the easiest number on any instant-swap site to inflate, because it is the union of upstream vendors' pair lists, not a promise. So on 20 November 2026, between 18:07 and 18:29 UTC, we asked our own quote engine the direct question. We drew 30 long-tail assets — a seeded random sample (seed 20261120) from every catalog asset that is neither a major nor a major stablecoin, landing on names from COMP and FIL to BANANAS31 and LADYS — and pushed each through three rails: 100 units into XMR, 1 XMR back out into the asset, and 100 units into USDT-TRC20. Ninety long-tail pairs, five major-pair controls, every quote fetched fresh through the measurement channel that skips the 8-second cache. The question is not whether we can list a coin. It is whether the engine will price it. (XMR to USDT exit)
The headline: zero 'pair not supported' in ninety tries. Across the 90 long-tail pairs, 47 returned a live quote, 40 were rejected on amount bounds — 'amount below minimum' or 'amount above maximum', always with exact numbers attached — and 3 failed with transient 5xx server errors (one 520, two 502). Not a single pair answered 404 'pair not supported'. The five controls all quoted: 100 USDT-TRC20 bought 0.22132 XMR across two routes, 0.01 BTC bought 1.309273 XMR, 1 ETH bought 3.621454 XMR, 1 XMR sold for 430.863838 USDT-TRC20, and 1 XMR bought 0.00746543 BTC. A quote engine that actually honors its catalog is rarer than the catalog itself — plenty of sites list hundreds of coins that only ever answer 'not supported' when you actually try. (XMR to USDT exit)
The real wall is the minimum, and it is priced in value, not units. Sixteen of the thirty assets refused a 100-unit deposit as 'below minimum'. Back-pricing each floor through the token's own quote on the other rail, the implied minimums cluster in a narrow band: SAND 331.31 units = $11.22, LUNC 247,754 = $10.70, SLP 27,085 = $11.23, DEP 11,085 = $11.56, PUNDIX 154.21 = $11.22, FIDA 851.92 = $10.86, PNK 1,928.06 = $13.15, LINEA-LNA 6,012.13 = $10.70, BANANAS31 2,187.20 = $13.11, NAFT 69,586.20 = $10.81, VTHO 25,724.52 = $10.66, BTTTC 36,793,658 = $11.26, LADYS 2,715,574,664 = $11.22. Thirteen tokens, thirteen wildly different unit counts, one band: roughly $10.66 to $13.16. The minimum is not a number of coins — it is a dollar floor wearing whatever costume the ticker demands.
Ceilings bite too, and they bite the expensive assets first. Three assets refused 100 units as 'above maximum': BTCB on BSC capped the corridor at 0.09457250 BTC — about $5,576 at that hour's spot — DPI at 75.39 units (about $2,975), COMP at 33.60 units (about $574). On the exit rail, DENT and MEW capped what 1 XMR could buy at 0.63180397 and 0.63155617 XMR respectively — roughly $282 per order. These are not the corridor-level limits documented in our earlier network sweeps; they are per-asset ceilings that sit on top, invisible until the engine tells you 'above maximum'. Anyone moving size through a wrapped or mid-cap asset should expect a different ceiling than the headline limit of the corridor.
Direction asymmetry is a dollar illusion. Entering XMR with 100 units of a long-tail asset quoted only 11 of 30; leaving XMR with 1 XMR quoted 27 of 30 (plus DENT and MEW hitting their ceilings). Nothing changed except which side holds the dollar value: 1 XMR was worth about $446 that hour and clears every floor, while 100 units of a $0.03 token is $3.40 — below any minimum by an order of magnitude. The lesson is unglamorous: before you type a unit count, multiply it out. The quote engine prices value; your input field accepts units. Sandbridge the two yourself.
Single-pool honesty on the long tail. Every one of the 90 long-tail quotes returned routes=1 with advantageBps of 0 — one pool priced it, no second quote existed to beat, and we refuse to dress that up. The majors behave differently: USDT-to-XMR ran two routes with a 997 bps gap at the $100 size (the pools disagree most exactly where the fixed fees weigh heaviest), BTC-to-XMR two routes at 108 bps, ETH-to-XMR two at 68, XMR-to-BTC two at 102, and the XMR-to-USDT exit ran a single route. The 'compared N routes' line is a real signal, and it is only real where liquidity depth justifies a second pool. (XMR to USDT exit)
The spread on a red day. That hour was not calm: Monero spot sat at $445.95 (CoinGecko, 18:05 UTC), well below its recent range, with BTC at $58,955 and ETH at $1,629.15. The two-sided XMR market priced the stress in: buying XMR with USDT implied $451.83 per Monero (+1.32% over spot), selling XMR for USDT implied $430.86 (−3.38% under it), and a 100 USDT round trip returned 95.36 — a 4.64% cost, versus the roughly 2% this corridor showed in calmer sessions. The BTC side was tighter: buying through 0.01 BTC implied $58,387 per Bitcoin (−0.96%), selling 1 XMR for BTC implied $440.12 per Monero (−1.31%). Wide spreads are not a defect to hide; they are what an aggregator honestly passes through when the underlying market is falling. (XMR to USDT exit)
The quote engine is a depeg detector, if you feed it the right pair. In the same minute, 100 USDS (an asset the catalog flags as a stablecoin) bought 0.20983982 XMR while 100 USDT bought 0.22132 — 5.19% less Monero for the identical face value. Back-priced against that hour's spot, one USDS was worth $0.9358 to the pools and one USDT $0.9870; USDQ priced the same at $0.9360. Whatever that 6% haircut is — market depeg, thin books, or risk premium — the engine passes it through verbatim, before you send anything. This is the quiet superpower of a multi-pool aggregator for stablecoin exits: you can watch two 'dollars' diverge in real time and pick the cheaper rail while the divergence is still on the screen. (XMR to USDT exit)
Reading errors is a skill, and the errors are generous. The measurement itself ran on them: a 400 body carrying minAmount and maxAmount is the pair handing you its working range for free — every floor and ceiling in this article was extracted that way, and the same trick works for anyone pointing a request at the quote endpoint. A 404 'pair not supported' means the corridor genuinely does not exist; stop retrying. The three 5xx responses (one 520, two 502, out of 95 calls) were transient upstream hiccups — worth one patient retry, not ten rapid ones, because the quote cache serves identical bodies for 8 seconds anyway. One caveat from our own catalog: an engine that answers everything cannot be taken for granted, so treat 'listed' as a hypothesis and let the bounds error be your proof.
Closing: the catalog passed, with fine print. Ninety long-tail pairs, zero unsupported, forty honest bounds rejections, three transient errors — and a wall of dollar-denominated minimums at $10.66-13.16 that no unit-based mental model survives. As of writing (20 November 2026, 18:29 UTC), the catalog listed 476 assets on 67 networks, 57 flagged stablecoins, six assets requiring an extra ID; 100 USDT-TRC20 quoted 0.22132 XMR across two routes, 1 XMR sold for 430.863838 USDT-TRC20, and spot Monero stood at $445.95. Quotes move; the live page governs. The method does not move: think in dollars, probe the bounds with a free 400, and never trust a catalog number you have not fed to the engine. (XMR to USDT exit)