Kukun opens permits API for real estate underwriting and AI

By AI, Created 11:51 PM UTC, May 29, 2026, /AGP/ – Kukun said June 2 it has launched the general availability version of its Permits API, giving lenders, insurers, real estate platforms and AI developers direct access to a normalized U.S. permit dataset spanning 750 million records. The company says the API is meant to make building permits a standard input for underwriting, valuation and property AI workflows.

Why it matters: - Kukun is trying to make building permits a default data layer for underwriting, valuation, insurance and real estate AI. - The API is built to expose a national permit signal that can show what has actually been done to a property before downstream risk or value changes show up elsewhere. - The launch expands access to a dataset Kukun says already supports production workflows at major banks, mortgage lenders, insurance carriers and real estate data platforms.

What happened: - Kukun announced the general availability of Kukun Permits API v1 on June 2, 2026. - The API provides programmatic access to 750 million normalized U.S. building permits across 57 million parcels and 2,400+ jurisdictions in all 50 states. - Kukun said the dataset is refreshed daily. - The API is aimed at lenders, insurers, valuation platforms, real estate technology companies and AI developers. - The product includes full-text search, multi-dimensional filtering and geo-distance queries on permit records tied to individual properties.

The details: - The permit corpus includes sub-permit enrichment that Kukun says surfaces about 40% more major residential activity than raw municipal feeds. - The API supports search across permit descriptions. - Users can filter by status, label, address, city, state, ZIP, county and jurisdiction. - The API returns daily-refreshed permit history at the parcel level. - The system replenishes cost data for roughly 50% of U.S. permits that arrive without contractor-declared values. - Kukun fills those gaps using its labeled taxonomy and proprietary ARVE renovation cost dataset. - The API includes a normalized taxonomy of 25+ project categories. - Technical access is based on standard REST, bearer-token authentication and JSON responses.

Between the lines: - Kukun is positioning permits as the earliest and most consistent signal in residential property data, which could improve underwriting and model accuracy when municipal records are fragmented. - The launch also broadens Kukun’s pitch beyond data licensing into infrastructure for AI developers building property workflows. - Kukun’s core thesis is that normalization and enrichment matter as much as raw permit volume, especially when separate municipal filings obscure the full project picture. - Raf Howery, Kukun’s CEO, said permits should be a standard input for any company underwriting, insuring, valuing or building AI for U.S. residential property. - Howery said Kukun has the only national enriched, query-ready permit dataset at this scale and that no other feed matches its sub-permit enrichment.

What’s next: - Kukun said a Model Context Protocol server is in active development. - The MCP layer is intended to let AI agents and frameworks such as Claude, ChatGPT, Cursor and LangChain call the property state graph natively. - Kukun says API access is available through api@mykukun.com and sales@mykukun.com, with a developer portal at portal.mykukun.com. - Kukun’s LinkedIn page is here.

The bottom line: - Kukun is turning permit data into infrastructure, not just a search tool, and betting that cleaner property records will become foundational to real estate underwriting and AI.