// CONSTRUCTION · B2B
Landing Live · In Build
Plain-English construction intelligence for NYC. Permits, ownership, building history — searchable like a conversation, not a database.
// The problem
Almost everything a NYC contractor needs to know about a building is in a public dataset somewhere — DOB permits, ACRIS ownership records, PLUTO land use, HPD violations, certificates of occupancy. The problem isn't access. The problem is that the interface is a 1998 government portal with a deep-link to a PDF.
Sourcing leads, vetting addresses, pulling history before a meeting — all of it is the same story: someone in a truck on Atlantic Ave trying to navigate a portal designed for a desktop in a Lower Manhattan office. The data is there. The interface treats it like a secret.
"Permits should be searchable the way you'd ask a person who knew. 'Show me everything within five blocks where the owner just changed.' That's the interface."
// What it does
Plain-English search
"Permits filed in Bushwick last quarter for facade work over $50k." The search bar talks to the datasets directly. No SQL, no field names, no cryptic codes — the model knows what you mean.
Building dossier
Type any NYC address. Get permit history, current ownership, violations, recent sales, comparable jobs — formatted for someone who needs to walk into a meeting in twenty minutes.
Lead scoring
Filter the city by owner change, permit type, violation count, age. Designed for contractors prospecting — not investors valuing.
Built for the field
Loads on a job-site connection. Searches return in under a second. No login wall for the basics. Built by someone who's watched supers try to use a desktop portal on a 4-inch screen.
// Preview
// Where it is
Done
In flight
Next
// Get involved
If you're a NYC contractor, supplier, or operator who wants the public data to actually feel public — get on the early list and you'll get a seat at the first cohort.
// Under the hood
Tech stack
Next.js + Postgres on Neon, vector layer over agent-rewritten summaries. Daily ETL across NYC DOB, ACRIS, and a few primary sources; queries route through Anthropic with citation grounding so every answer is reproducible.
Orchestration via Drako DevOS →
GitHub
Closed-source — proprietary data ingestion. Generic search/grounding patterns will land in the OSS Drako DevOS release.
— COMING WITH OSS RELEASE
// Other projects