Augea

I've been building Augea (augea.io), a tool that compares what it actually costs to buy crypto across exchanges, countries, and payment methods.

The hard part wasn't scraping fees. The hard part was deciding what to do when the data is uncertain.

Every comparison site I looked at does the same thing: pick a number, show it confidently, link to an affiliate. If the fee changed yesterday, if the exchange hides part of the spread, if the payment method surcharge depends on your country, you'd never know from the UI. The number just appears, clean and final.

That felt dishonest. So I built Augea around a different rule: every number has to carry its provenance. When was it captured. Which exchange. What verification level. If we don't know something, the UI says "unknown" and explains why, instead of guessing.

Three things I didn't expect:

1. "Unknown" is a feature, not a bug. Users trust a tool that admits what it doesn't know more than one that pretends to know everything. The confidence signal matters more than the number itself.

2. Truth-state beats ranking. I spent weeks trying to rank exchanges "best to worst." It was the wrong frame. What users actually need is: "here's the cost range, here's how fresh the data is, here's what we couldn't verify." Ranking implies certainty we don't have.

3. The moat is boring. Provenance infrastructure, snapshot pipelines, tier-gating rankable vs unrankable data. None of it is glamorous. But it's the only thing that makes the output defensible when someone asks "where did that number come from."

The product is still early. Coverage is thin outside a handful of countries. The UI is English-only. But the architecture is built so that when we add a new exchange or country, the truth-state rules apply automatically, and we never have to retrofit honesty.

Curious if anyone else here has shipped something where saying "I don't know" in the UI was a deliberate design choice. It's uncomfortable at first, then it becomes the thing users remember.

— augea.io