Every yield number with its source attached. And its risk record.
PILOT YieldScan scans public DeFi and staking yield sources and hands you one report: what each source was paying at the moment of the scan, where each number came from, and the risk signals around it, how old the protocol is, whether audits are published, which way TVL is trending, whether there is a past incident on record. Then it stops. No recommendation, no execution, no custody, no keys. A report you can check, not a pitch you have to trust.
Get a source-attributed yield snapshot See a sample report
THE PROBLEM
Yield information in crypto arrives as screenshots. An APY number floats through a feed with no source, no timestamp, and no mention that the protocol paying it is four months old, unaudited, bleeding TVL, or carrying an exploit on its record. The aggregator boards are better, they give you a live number, but the number arrives naked: provenance in one tab, audits in another, incident history in a third, if you think to look at all. And the people who do package it all up for you usually want something: they are selling the protocol, farming a referral, or running a bot they want your funds inside. October 2025 is the standing reminder of how that ends, a $19B liquidation cascade and a headline-yield stablecoin depegging to $0.65, with most holders discovering the risk when the market did, not before.
HOW IT WORKS
You name an asset. YieldScan queries the wired public yield sources and builds one table: each venue, the yield it was paying at scan time, and four risk-signal columns against every row, protocol age, audit status (auditor names and dates, when published ones are surfaced), TVL trend over the trailing window, and past-incident flags (recorded exploits, depegs, halts, fund-loss events). Every single number carries its source and its observed-at timestamp; a number that cannot be attributed is not reported at all. The whole report is stamped as a point-in-time observation, because yields move and vanish, sometimes between the scan and your screen.
Three things it will never do, by construction, not by policy: it has no recommendation field in its output schema, so it cannot tell you where to put money even badly; it has no execution path, so it cannot sign, submit, or simulate a transaction; and it takes no keys, funds, or account credentials, so there is nothing of yours for it to hold. Read-only public sources in, attributed report out, done.
WHAT YOU GET
For every scan, one report you can cite: - the yields, per venue, as observed at scan time, each with source and timestamp, - the risk signals, protocol age, audit status, TVL trend, incident flags, each attributed, - the comparisons, net-of-cost arithmetic between venues, shown as observations, never as a recommendation, - the caveats on the face of it, the point-in-time stamp, the source-coverage list, and the false-negative disclosure printed on every report: the absence of a risk flag is NOT evidence of safety, it means the wired sources surfaced nothing.
Signals, not ratings. YieldScan will not compress four signals into a safety score, because a single score is exactly the kind of number that gets screenshotted without its caveats.
WHO IT'S FOR
DeFi and staking participants who actually do their own research and are tired of doing it across nine tabs. Crypto writers, analysts, and small desks who need a yield table they can publish or put in front of a committee with the caveats attached. If you want a tool to tell you what to buy, this is not it and will not become it. If you want a bot to move your funds to the best yield, that is the opposite of this product, and the refusal is the point.
PRICING (the ladder)
We price the ladder, not a single number, and every number below is a hypothesis we validate, not a commitment. - Free snapshot, a one-off attributed, risk-annotated yield report for an asset you name. The free tier is the trust argument: check our numbers against our sources. - Subscription, scheduled scans, change alerts when a tracked yield or risk signal moves, and scan history, so point-in-time observations become a citable series. - Analyst, multi-asset coverage, exportable attributed tables, API read access for desks and publications.
No tier ever includes advice, execution, or custody. IP is licensed, never assigned. The only money step in this product is the subscription charge itself, and it is gated and explicitly confirmed before anything is charged, nothing charges silently.
THE PROOF (dogfood)
Honesty first: AYA holds no yield positions and stakes nothing, so we cannot show you our own portfolio running on this, and we will not pretend to. What we dogfood is the discipline the product is made of: every capability AYA ships terminates as a verification receipt on disk, every claim carries its evidence path, and an undeclared green check is treated as a defect. That is the same contract YieldScan offers you, every number attributed, every limit declared. The build left its own receipts too: the deliverable pattern this product runs on was scored by our quality gate (92.8, grade A) and both of the capabilities it composes were verified on disk before this page was written.
HONEST NOTE
We would rather under-promise. The YieldScan deliverable pattern exists, is structurally verified, composes two real capabilities (a read-only vault-data fetch and a comparison step), and every referenced pattern resolves on disk. What it has NOT done yet: no live yield scan has ever run, for a customer or for us, because no yield-source connector is credentialed yet and none of the risk-signal feeds (TVL, audit registry, incident database) are wired. The risk-signal taxonomy exists as authored data awaiting those feeds. The sample report is a structural sample and is labeled as one. The first live test is the natural next step: wire one yield source and one TVL feed read-only, run one scan, file the receipt. Until then this page describes what the product is built to do, not what it has done, and we will keep that distinction visible everywhere, because a yield tool that blurs the line between built and proven has no business annotating anyone else's risk.
Reminder, because it belongs at the bottom too: informational only, not investment advice, yields are point-in-time observations that can vanish, risk annotations are signals not ratings, and the absence of a flag is not safety.
Get a source-attributed yield snapshot
*This page is a specification. The capability it describes is not built yet, and nothing here is a claim that it runs today.*