European Union crypto regulation changed fundamentally with MiCA — the Markets in Crypto-Assets Regulation. But the raw data behind it is deceptively messy. AssayDesk is a small, focused tool that turns five ESMA-published CSV registers into something you can actually reason about — without inventing facts along the way. Here's how it works. Where the data comes from ESMA — the European Securities and Markets Authority — publishes five MiCA registers as CSV files: Crypto-asset service providers (CASPs) E-money token (EMT) issuers Asset-referenced token (ART) issuers Crypto-asset white papers Entities flagged as non-compliant by national competent authorities AssayDesk fetches those files on a daily schedule. ESMA itself republishes roughly weekly, so most days the file is byte-identical and nothing changes. Every fetched file is archived with its SHA-256 hash before parsing. That means any figure on the site can be traced back to the exact bytes ESMA actually served — you're not taking anyone's word for it. What "as published" actually means The register tables store ESMA's values verbatim — including truncated names, malformed dates, duplicated URLs, and blank fields. Nothing is silently repaired on the way in. This is a deliberate design choice: if a value looks wrong here, it looks wrong in the source — and that is information in itself. A messy official register is itself a finding about how the data is being managed. One consequence worth knowing: ESMA publishes one CASP row per authorised service, not per firm. So a single company appears once for each service it may provide. The sidebar counts entities; the register page shows the service rows. What curation changes (and what it never does) Deterministic rules produce a display-ready value next to the raw one — a repaired link, a normalised date, a trimmed name. The rules are versioned, and each correction records which rule produced it. The original value is always kept and always shown alongside. A correction is an annotation, never a replacement. A weekly model review may add observations about suspicious records. It never writes a value — and any model output is labelled unverified until a human confirms it. What the change history can and cannot assert The change history records what AssayDesk's sync observed between two published files: a record appeared a field changed a withdrawal date was published a record vanished These are observations about the file — not statements about the firm. The distinction matters: A record disappearing from a CSV is not evidence that an authorisation was withdrawn. ESMA states withdrawals as a date column; disappearances are usually publication artefacts. Absence from the register is not by itself proof that an entity is unauthorised — you should check the national register of the relevant member state. Why this matters Most crypto-regulation dashboards quietly "fix" bad data and never tell you. AssayDesk does the opposite: it keeps the raw bytes, shows you exactly what it changed and why, and only annotates — never silently rewrites. That's the real product: a trustworthy lens on an official but messy dataset, where provenance and honesty about uncertainty are the headline features, not an afterthought. Start from the overview, or read the data-quality report.