The same work happens every week and a person does it
Collecting, reformatting, checking, filing, posting, reporting. It is not hard work, which is exactly why it never gets automated: it is always cheaper this week to just do it again.
Automation that runs the boring half of a business.
The reason eleven products can be carried by a studio this size is that the identical parts of eleven launches are done once, by a system, and reviewed by a person before anything leaves. That system is not for sale. The way it is built is what we would build for you.
What people arrive with
The problems that bring somebody here.
You tried a model and could not trust the output
A model that is right most of the time is unusable in a process that has to be right every time. The fix is almost never a better prompt; it is a schema the output has to satisfy, a gate it has to pass, and a human review on the one step where being wrong is expensive.
An automation that fires twice is worse than none
Retries, crashes and double triggers are normal. An automation that posts twice, charges twice or emails twice because of one is not an edge case - it is the default, unless identity is designed in.
What the work includes
Including the parts nobody puts on a services page.
Scope is where a project is decided, so this list is specific rather than reassuring.
The whole loop, not one step
Research, build, test, review, publish, measure, improve. Automating one step of a seven-step process moves the bottleneck rather than removing it, and the measure step is the one most often left out and most worth having.
Jobs that cannot run twice
Work is identified by a deterministic hash of its own inputs, so a duplicate trigger, a retry or a crash-and-restart resolves to the same identity. Retrying is always safe, which is the property that makes automation restartable at all.
Gates, and a person on the last one
Automated checks run first and fail closed. Publishing is held behind human approval by design, not because the automation is unfinished - our own system has published nothing without a person saying so.
Evidence instead of a dashboard
Every run leaves an artefact: what it decided, on what input, and why it was allowed to. That is what makes an automation debuggable six months later, and what lets you answer a question about a decision nobody remembers making.
Integrations, with their failure modes written down
A third-party interface will be slow, rate-limited, down, or will change a field. Each integration gets a stated behaviour for each of those, because the interesting part of an integration is never the happy path.
On-device and local models where they fit
Not everything needs a hosted model. One of our products runs pose detection natively on the phone; another generates images against a GPU in the studio. Where the work can stay on the device it is cheaper, more private, and still works on a train.
How it runs
The same loop our own products go round.
Not a process diagram written for this page. It is the production loop the studio already runs, which is why it has a measure step and a human gate in it.
- 01ResearchSignals and a per-product idea bank, scored for relevance before anything is made.
- 02DesignThe product decides its own voice, claims and boundaries. The system fills them in.
- 03BuildAssets, narration and captions assembled from each product’s own material.
- 04LocaliseEvery string in every language, with parity between catalogues checked rather than assumed.
- 05TestAutomated gates, then a person. Nothing reaches a queue unreviewed.
- 06PublishPer-platform packaging, with publishing deliberately held behind human approval.
- 07MeasureRelease and performance signals return to where the work was made.
- 08LearnWhat worked on one product is readable by the next one that asks.
- 09ImproveThe fix lands in the shared system, not in one product’s copy of it.
Evidence
Our own products, not a client list.
Everything shown as evidence on these pages is a NOVEMADE product - our idea, our money, our name on the listing. There is no client logo wall here, no testimonials and no case-study metrics, because there have been no clients to take them from. What there is instead is software you can open.
Sprint analysis, training and nutrition in one app, with the form analysis on the phone.
iOS · AndroidTry it
Design a tattoo, turn it into a working stencil, and see it on your own skin first.
iOS · AndroidTry it
Check it yourself
A walkthrough, not an analysis. Nothing is uploaded and nothing is computed in your browser — you step through the stages the app performs, and the report at the end is a real screen captured from SPRIVO rather than a number invented for this page.
Goes further
Eleven products, one production system, and the order we read evidence in
The expensive parts of a launch are identical every time. Sharing them is the obvious half; knowing which document to believe is the half that keeps going wrong.
Where this stops
What we do not do.
No credentials, endpoints, machine names or internal paths appear anywhere on this site, and none would appear in your project either. We also do not sell the studio’s own production system, and we do not agree to automate a process nobody has yet written down - the writing down is the first half of the work and it is not optional.
Also from the studio
The other services.
- Mobile app developmentiOS and Android development by an independent product studio that ships its own apps. Eleven products, seven live on a store, each one playable in your browser before you ask us anything.
- Custom softwareCustom business software, internal tools, dashboards and APIs, built by a product studio that maintains everything it ships.
- Web developmentHigh-performance Next.js websites from a studio whose own site is served in eight languages with no third-party scripts and a strict content-security policy.
- E-commerce developmentCommerce and payment architecture from a studio that has shipped subscriptions, entitlements and verified purchases across its own products - and has not shipped a storefront.
- Product developmentEnd-to-end product development - strategy, design, engineering, release and iteration - from a studio with eleven products of its own and seven of them live on a store.
Collaborate
We are not for hire. We are open to the right partnership.
Publishing, partnerships, investment, or a technical conversation about how we build. Tell us which one and we will skip the rest of the form.