Software businesses,
built and run.

Every real estate site shows you the kitchen. None tell you the reservoir drops forty feet in November, that the dock needs federal approval, or what the first year actually costs. This one does.

Live ingestion from U.S. Geological Survey sensors. True shoreline boundaries from the National Hydrography Dataset. Deterministic scoring across five axes with the formula published on the site.

Every listing site sorts by price. This one sorts by what you will spend over five years — fuel or charging at your local rates, depreciation, known failure modes, and whether the dealer has a history of bait pricing.

Cost-of-ownership scoring built on federal public-domain sources: NHTSA recall and NCAP crash data, NREL charging stations and EIA energy rates. Listing data source under review.

Most acquisition tools start after you have a deal. This one tells you which listings are worth a phone call, and why the ones that look good usually are not.

Scores a listing against asking price, seller discretionary earnings, asset coverage, and goodwill multiple. Confidence is a progress mechanic: the report tells you exactly which documents would move a verdict.

A scoring system for business ideas. It runs candidates through fixed gates and kills the ones that do not clear them — including several we wanted to like.

Three tracks with separate gates and weights. It has blocked our own ideas, which is the point of building it.

Research and signal generation that runs without supervision.

GA4 Fix, AD Sync Fix, and Fail2ban for journald. Cheap to build, cheap to kill. Every one had a pass mark set before launch.

Deliberately small. The point is a real answer about demand, not revenue.

A controlled remote interface to the server. Twelve tools behind four independent security layers, every one of them tested by trying to break it.

Path allowlist that resolves before it checks, so symlinks cannot escape. Sixteen refused command patterns. Full audit log.