HomeCompaniesFeldy

Field-First AI for home services

Feldy is field-first AI for home service businesses helping teams capture jobsites, create take-offs & estimates, and documentation. With Feldy, teams capture the physical world using photos, videos, AI jobsite walkthroughs, and 3D scans using your mobile phone. It turns field information into takeoffs & estimates, visual designs, and project documentation. Feldy works across home services such as roofing, exteriors, remodeling, painting, cleaning, landscaping, handymen, and other home services. We’re building a new way for more than 2 million U.S. home service businesses operating in an $800B+ market to capture, understand, and complete work in the field. Feldy is available on iOS, Android, and the web.
Active Founders
Zayne Sagar
Zayne Sagar
Founder
Zayne is the co-founder and CEO of Feldy. He completed an Applied Math/CS and Philosophy bachelor's degree from UC Berkeley, and attended graduate school working on topics around computational science methods. He is a fan of building physics-grounded world models using multi-modal inputs.
Shelby Bons
Shelby Bons
Founder
Shelby is the co-founder and CTO of Feldy. She has expertise in software engineering and ML Physical Systems modeling, developed at leading climate-centric organizations like Sunrun and Voltaiq. Before that, she completed a Materials Science and Engineering degree at the University of California, Berkeley. Prior to founding Feldy, Shelby was headed for a Theoretical Physics PhD career track.
Company Launches
Feldy: Field-first AI for home services and contractors.
See original launch post

TL;DR 👋

Feldy is field-first AI for home service companies. It unifies jobsite capture, 3D building scans, estimates, and field documentation in one simple app.

We’ve grown to over 200 paying customers on the web over the last few months, and are now launching the mobile app so field teams can use the full product on site.

Zayne and Shelby are construction software veterans with a combined 10+ years building computer vision and spatial AI tools.

Meet Feldy:

https://www.youtube.com/watch?v=W0lEF7isKg4

The Problem:

The biggest problem in residential home services in 2026 is still remarkably simple: estimates do not go out fast enough to homeowners.

An estimate is the quote a homeowner receives explaining what work will be done and how much it will cost. Yet, if you call a window contractor, or painter, it can still take two or three days, often longer, to receive one. Over the last 8 months, my home became a test bed for remodeling projects, and this is one of the few truths that persisted. This problem appears in nearly every trade we encountered.

So why so long?

The surprising part is that many of these estimates are not inherently complicated.

Consider a typical window and door contractor. A sales representative drives to the property, walks through the home, identifies every window and door, photographs existing conditions, and asks discovery questions:

  • Do you want to replace the trim?
  • Should the opening change?

Capturing this information can easily consume most of a one- or two-hour appointment. By the time the rep understands the property and records everything the office needs, there is very little time left to actually sell the project. This is the primary reason quotes arrive days later, or require a second visit.

Furthermore, existing software makes the problem worse.

Most contractor tools have always been structured like digital forms. The software records what the estimator has already figured out, but it does very little to help figure out the job.

This is what led us to ask: what does this contractor world look like if the software recorded and figured out the job at once? Well, that’s Feldy.

uploaded image

So how do we solve this?

Instead of requiring the person in the field to manually observe and organize everything, Feldy uses in-house computer vision and vision-language models to perform much of this low-value work: necessary technical work that consumes the majority of the job visit time, but creates little visible value for the homeowner.

The contractor captures the property via photos/videos, lidar scans, or job walkthroughs (video and audio are combined). Feldy then interprets the capture and helps turn it into an estimate amongst a host of other workflow items such as inspection reports, proposals, materials lists, etc.

The entire thesis behind Feldy can be summed up as:

How do we remove the time-consuming thinking required from the estimator that creates little visible value for the homeowner?

uploaded image

The Scale of the bottleneck

Angi’s list has estimated that American households undertake more than 700 million home projects annually.

Using 600 million estimates (aggressive case):

  • 600M estimates * 3.5 hours = 2.1 billion total labor hours
  • 50% spent on measuring, documenting, takeoff, and estimate construction = 1.05 billion hours
  • At $45–50 fully loaded cost = $47–53 billion in annual labor capacity tied up in that work

If Feldy reclaims 1–1.25 hours per estimate:

  • 750 million hours returned
  • Roughly $27–37 billion in productive capacity per year.

For the average home-services company, this is equivalent to reclaiming $50,000–$100,000+ in capacity each year, capacity that can be converted directly into more estimates, lower overhead, and higher revenue.

The Team

Zayne and Shelby met at UC Berkeley’s engineering school nearly 10 years ago. For the last three years, they have worked side by side building technology for the construction and home-services industries.

Together, they previously took a construction-software company from zero to several million dollars in ARR.

The pair is more excited than ever to build the next generation of software for home-service businesses.

The ask

If you know anyone who works in home services, tell them about Feldy and ask them to download the app!

Search Feldy AI on IOS and Android app stores.

Previous Launches
Aether is AI design and workflow software for solar installers.
Feldy
Founded:2024
Batch:Winter 2024
Team Size:15
Status:
Active
Location:San Francisco
Primary Partner:Gustaf Alstromer