What we build
CHEEE is an independent AI-product startup founded in late 2025 by JR and SP. We build across two connected areas: practical applications for everyday work, and browser games designed around clear mechanics and engaging play. Product design, frontend engineering and AI-assisted development connect both.
Useful applications. Memorable play.
Tools for everyday work
Our applications address repeated tasks, information that needs organising, and handoffs that are easy to lose. These jobs often sit between larger systems and everyday work. We design focused browser tools that help someone see what is ready, what needs attention and what happens next.
We shape the first version around one clear job instead of adding features for their own sake. That could mean gathering information, preparing a draft, or showing a person which items need a closer look. AI can assist with a defined part of the process, while review and final decisions remain clear.
Before building, we want to understand who uses the tool, what information they already have and what they do when something does not fit the usual pattern. Those details help keep the workflow useful in ordinary day-to-day conditions.
- A repeated requestCollect the details in one place so someone can check them before work begins.
- A handoff between peopleMake the next step and the information it needs easier to follow.
- Notes that need a reviewAn AI helper could organise a first draft for a person to check.
Games and learning experiences
Our games bring together browser play, creative coding and strong mechanics. Some are suited to campaign or learning goals; others are made for the pleasure of movement, timing and discovery. We start with what a player can do, then shape the feedback and rules around it.
A possible experience might introduce an idea, invite someone to make a choice, then show how the choice changes the result. It should be clear what to do next and how the activity connects to the campaign goal.
- A choice to practiseA short game around a decision your campaign needs to explain.
- An idea to tryAn interactive learning activity that lets people see what changes when they make a choice.
- A campaign to take part inA browser experience with a clear activity and an understandable finish.
AI in the making.
Product value in the result.
Claude, Codex and OpenCode help us move from product decisions into working software. We use them for code exploration, prototyping, frontend implementation, debugging and iteration, with JR and SP directing and reviewing the work.
AI-assisted development describes how a product is built. AI-powered functionality describes what it does for someone using it. We make the second claim only when a live model feature is implemented and verifiable.
-
Development with Claude
Claude is part of our AI-assisted development workflow, alongside Codex and OpenCode. It supports implementation and problem solving; people remain responsible for the product, code review and final quality.
-
Inside an application
A model API is useful when it performs a defined job, such as organising notes or extracting information. Runtime AI is a deliberate product feature, not an automatic property of every app.
-
Across both tracks
We connect the product question to a working interface, check how it behaves and refine it against the job it needs to do.
Bring the task.
We can work out the first step.
Describe how the work happens today, where it gets difficult and who is involved. For a campaign, tell us who it is for and what you want people to learn or do. You do not need a finished brief: a rough outline, an example of the task or a question you are still working through is enough to begin.
How we approach a projectDescribe the task or campaign.
If you have an example of the task, a rough brief or notes about the campaign audience, include them. If the idea is still taking shape, tell us what feels difficult or what you want people to understand.