of the Day


Lens of the year award


Go download the app

FieldDay lets anyone create real-time,
vision-AI applications, right on their phone.

Its all-in-one design allows anyone, from hardware tinkerers to creators of AR filters, to build computer vision-enriched apps fully customised to their domain expertise.

From data-collection to deployment, no need to write code or know about ML.

The Problem

Vision AI has quietly found its way in many of the products we know and love. Equally, creative and developer tools have built out support for integrating computer vision models in creators' workflows.

Building models however remains an expert task, often requiring a dedicated machine learning engineer to work alongside the creative team for weeks, making it unfeasible for creators' agile workflows with quick turnarounds.

Even if a model does get built in such a fast-paced environment, real-world testing is often a last-minute task, letting the model fall flat in the real world.

The Solution

With FieldDay creators and tinkerers themselves can build models tailored to their needs without involving a dedicated engineer.

Building happens in a quick, and iterative loop from data-collection, through to training, and testing to eventual deployment to almost any platform.

Built with FieldDay

Our community

After we announced FieldDay, we found builders from all types of backgrounds and industries who were interested in the idea of creating truly custom machine learning. Below you see a summary of some of the things people have built with FieldDay.

How we built it

FieldDay is the culmination of years of work in participatory, data-first machine learning. We have taken everything we know about making approachable tools to build ML.

Smart, and delightful.

Interaction design

Competitors in this category are highly technical, and often hard to use.

If we wanted to make FieldDay appealing to a non-technical audience we knew that the app had to feel fun and powerful.

Minimally lovable.

Brand Design

The brand was developed to be effective across many channels and applications while at the same time being as minimal as possible.

We put our emphasis on an identity that had a root in the technical foundation of FieldDay ('the learning loop'), leveraging radial gradients as the primary texture to reinforce the iterative notion.

Driven by simplicity.

Technology and AI

FieldDay's technology stack was designed for simplicity first. Wherever we could assist the user and help them avoid mistakes we designed technology around it to make that happen.

Beyond just making machine learning approachable, the other technical hurdle was dealing with the mobile platform itself in a reliable way.