Why we build offline-first, not cloud-first

17 July 2026 · 6 min · Engineering

A classroom with a single device and no visible cabling

Most software is designed for a world with stable power and fast internet. That world is not where many schools and small businesses live. So we design for the world that exists.

The assumption that breaks

Cloud-first products assume the network is always there and always cheap. When it is not, the product simply does not work, and the people who needed it most are the first to be excluded. We invert the order. The product must work with the network off. Connectivity becomes a bonus, not a requirement.

It is also cheaper and safer

Offline-first removes recurring connectivity cost and removes the risk of sending student or customer data to a third party. For a school, that is a safeguarding decision. For a business, it is a continuity decision. Both are reasons enough on their own.

None of this is nostalgia. It is engineering for the actual constraint. The result is a system that gets better when the grid is stable and simply keeps working when it is not.