04 · MVP

AI MVP development: your first product version in 8 to 16 weeks

AI MVP development in eight to sixteen weeks: your first app version in users' hands, with a capture flow and filming guidance, the on-device model from the proof of concept, results and history screens, a minimal backend for upload and analytics, and an internal or TestFlight build. Priced from the PoC.

Timeline8 to 16 weeksPriceEstimated from the proof of concept
The first version assembling inside a phone: camera feed, pose layer and app interface, built for 20 pilots

Who it is for

Is this the right step for you?

  • The proof of concept passed and you need pilot users or first paying customers on the feature.
  • You are raising on real usage and need the app, not the demo.
  • You need the app built around the model, whether you have developers of your own or not.

What you get

First product version: the deliverables

  1. 01

    A capture flow with filming guidance: distance, framing and lighting checks before the user records, because this is where measurement apps fail first.

  2. 02

    The on-device model from the PoC, wired to the screens.

  3. 03

    Results and history screens.

  4. 04

    A minimal backend for upload and analytics.

  5. 05

    A TestFlight or internal build on the agreed devices.

  6. 06

    All code and IP, in your repositories.

How it works

From first message to done

  1. 1

    Week 1: scope, device list and the capture flow design.

  2. 2

    Sprints of two weeks, each ending in a build you can install.

  3. 3

    Pilot users' recordings go back into the model as they arrive.

  4. 4

    Handover: documentation and a walkthrough for whoever maintains the app next.

Timeline8 to 16 weeks
PriceEstimated from the proof of concept

The team is an AI tech lead, an AI engineer, a mobile or web engineer and a project manager.

Done when

PoC accuracy holds on pilot users' recordings, and the build runs on the agreed devices.

Compared with the alternative

An app agency without vision experience builds the screens and treats the model as a black box. The capture flow, the device spread and the model's failure cases are then nobody's job. Here the same team that trained the model builds the flow that feeds it.

Questions

What teams ask about first product version

Whichever your pilot users carry. The second platform is a product build item.

Yes. We build the capture flow and the model integration in your visual style.

Yours. The build ships from your account.

How we work

What you can count on

Will it fit what our developers already built?

Yes. The model ships as a module your team plugs in, and the backend is minimal by design.

What if we pivot?

The model and pipeline are modular and retrainable, and you own the code, so a pivot does not restart the work.

How does this compare with hiring?

It costs less than US engineers, and it can pause between funding stages without laying anyone off.

Next step

Send what you have. Verdict in a day.

Book a scoping call if your PoC is done, with us or with someone else. Otherwise start with the free assessment.

  1. 1
    You send a spec and a few recordings

    A paragraph on the feature, a link to clips or sensor logs, the platforms. NDA first if you need one.

  2. 2
    An engineer reviews it

    The CTO or a domain expert goes through the measurement, the conditions and the platforms.

  3. 3
    You get the verdict within a day

    Whether the feature can be measured, the risks we see, and the recommended next step, in writing.