catnapstudio

AI software& mobile apps,drawn upand built.

catnapstudio is a product studio. We design and engineer AI SaaS products and iOS and Android apps, end to end: for founders starting from a blank sheet, and for teams adding AI to software that already exists.

A
InterfaceScreens, flows and the design system
B
DeviceNative iOS and Android, or cross-platform
C
ApplicationAPIs, auth, payments, data and backend
D
IntelligenceLLM features, agents, retrieval, evals
E
InfrastructureCloud, monitoring, releases and scale

What we build

S-101 · Scale 1:1

Five items of work. Take one, or take the whole schedule and we draw up the product from the first screen to the servers underneath it.

Schedule of work
MarkItemScopeYou receive
S1ACDEAI SaaS productsNew products built around language models: assistants, agents, retrieval over your own data, structured extraction. Measured against real examples before they reach users.A shipped web product and the codebase behind it
S2CDAI in products you already runScoped AI features added inside existing software: search, assistants, automation, summaries. Built into your stack and your codebase, not bolted on beside it.Production features, documented and handed over to your team
S3ABCMobile appsiOS and Android apps, native or cross-platform. Offline behaviour, notifications, subscriptions and on-device intelligence where it earns its place.Apps live on the App Store and Google Play, plus the backend they need
S4AProduct designFlows, interface and prototypes, drawn before anything is built and put in front of people while changes are still cheap.A design system and clickable prototypes
S5ELaunch and iterationAnalytics, monitoring and release pipelines, then the next revision after the first one ships.A product that keeps moving after launch day

Two details, drawn large

D-201 · Enlarged

Most of what we build comes down to one of these two drawings, or both on the same sheet.

Detail 1 · Layer D, in section4:1
A request travels from the user to the app, which calls retrieval over your data and a model, passes the answer through evaluation, and returns a response.USERAPPRETRIEVALDATAMODELEVALSRESPONSEASKCONTEXTLOGGED · MEASUREDA request travels from the user to the app, which calls retrieval over your data and a model, passes the answer through evaluation, and returns a response.USERAPPRETRIEVALDATAMODELEVALSASKCONTEXTRESPONSELOGGED · MEASURED

NoteEvery model call in the loop is logged and measured, so the feature can be tested and improved like the rest of the code.

Detail 2 · Layers A–B, in section2:1
A phone running a chat interface, sectioned into four layers: interface, state and offline, sync, and the API behind it.iOS · ANDROIDINTERFACESCREENS · MOTION · A11YSTATE & OFFLINEWORKS WITHOUT SIGNALSYNC & PUSHNOTIFICATIONS · PAYMENTSAPI & AIA phone running a chat interface, sectioned into four layers: interface, state and offline, sync, and the API behind it.iOS · ANDROIDINTERFACESCREENS · A11YSTATE & OFFLINENO SIGNAL NEEDEDSYNC & PUSHPUSH · PAYMENTSAPI & AI

NoteThe app keeps working without signal and catches up when it reconnects. The intelligence lives behind the API, or on the device when it has to.

Sequence of works

P-301 · Not to scale

The same five steps for a new product or a new feature. Only the length of the line changes.

  1. Discover

    The problem, the people and the constraints. We agree what the first version has to prove.

  2. Draw

    Flows, interface and architecture on paper first, where changing your mind costs nothing.

  3. Build

    Short cycles and working builds you can open on your own phone as they happen.

  4. Ship

    Web launch, App Store and Google Play review, with monitoring in place from day one.

  5. Iterate

    Watch what people actually do, then draw the next revision.

General notes

N-401 · Read before building

Who we draw for, and how the work is agreed.

  1. Starting from a blank sheet

    Founders with an idea and no team yet: we take it from first sketch to a product in people’s hands, and stay on for the revisions.

  2. Adding to an existing drawing

    Teams with a product already running: we add AI where it earns its place, inside your codebase and alongside your engineers.

  3. All dimensions are negotiable

    Scope, timeline and budget are drawn up together before any work starts. Nothing is fixed until you sign the sheet.

  4. Do not scale drawing

    If something here is unclear, do not guess. Write to us and ask.

Bring us the first sketch.

X-501 · For construction

A napkin, a deck, or a product that needs AI in it. Tell us what you want built and we reply with questions, then a drawing.

[email protected]
Start a project
Revision history
RevDateDescription
A2026Sheet set issued · © 2026 catnapstudio