We are building the resolved record of American real estate.
The public record already contains the answers. It just does not contain them in a form anybody can use, and nothing in it cites itself. That gap is the company.
One machine, three states, no aggregator.
Mimir ingests 127+ free government datasets and resolves them into one canonical record per building and per person, then links those records into a control graph with an evidence chain on every edge. It runs self-hosted on commodity hardware and can be rebuilt from the raw sources.
We are a small team in New York. The name is from the Norse well of knowledge — you had to give something up to drink from it, which felt about right for public records.
Three positions everything follows from.
- An answer without a source is a rumour.
- Property data has been sold as confident assertion for decades. The moment a number matters — a credit file, a court, a committee — the question is always the same: how do you know? A product that cannot answer that is selling comfort, not information.
- Determinism is a feature, not a limitation.
- It would be easier to cluster records with a model and ship a score. It would also mean answers drift between runs and nobody can reconstruct why two records merged. We gave that up on purpose.
- The limits belong on the marketing site.
- Coverage is three states. Precision and recall are not published. New Jersey has no control graph. Those facts are on this website because a buyer will find them anyway, and finding them from us is worth more than the version where they are hidden.
Get in touch
Design partners, data licensing, press or anything else —sean@mimirgrp.com.
Mimir Group · New York
Bring a building you already researched by hand.
The fastest way to judge the system is to watch it re-derive something that took you a day, with a citation on every step.