Restb.ai

Restb.ai provides computer vision and AI tools that analyse real estate photos and turn what they show into structured data. The company says more than 1,000,000 property photos are uploaded each day in the US alone, and its software extracts details such as room type, materials, property condition and photo quality from them.

The product is a suite of solutions. Image tagging classifies photos by room type and identifies interior and exterior features and architectural styles. Property condition analyses a property's condition and quality from its imagery. Visual similarities finds the most similar properties and images within a market. Image captions auto-populate image alt tags. Comparable properties adds photo-based quality and condition scores to comparable sales. Photo compliance flags images and videos that break an MLS, website or report's guidelines. Property descriptions generates listing descriptions. Watermark detection checks uploads for added watermarks, logos and text. Duplicate detection finds exact and near-duplicate images and listings. Complexity assessment estimates how difficult an appraisal will be.

The page says the software is aimed at anyone who works with property imagery. It names five groups: AVM, iBuyer and investor companies; appraisal and inspection firms; multiple listing services; property search and marketing businesses; and insurance companies. Each group has its own uses, such as enhancing valuations for AVMs, automating report quality checks for appraisers, and auto-populating listings for MLSs.

According to the site, the data is standardised. Human descriptions of condition vary, and the company says its AI scores every property in the same way, so the output is consistent across a market.

The site lists case studies with claims such as a 46% increase in Google web traffic for a real estate portal using image captions, and a 9.2% reduction in error rate for an AVM using property condition models. It also says the company has built computer vision and machine learning models for the industry since 2015.

Pricing is not published. The pricing page describes the offer as scalable solutions and says the company aims to be more cost-efficient than building an internal AI team, but it gives no plan prices. Customers can book a demo or speak with an expert through the site.