A financial analytics and portfolio management platform that turns scattered account data into decisions people can actually act on.
- React
- Node.js
- Python
- PostgreSQL
Lumina
Brand-safe creative at campaign speed
An AI creative studio where marketing teams generate campaign imagery that stays inside their brand rules, with every asset carrying a record of exactly how it was made.
- Industry
- Marketing, Business
- Year
- 2026
- Screens
- 09
Overview
About the project
Lumina is a concept product for in-house brand teams asked to produce more creative than they can hand-make, without letting quality or consistency slip. It turns a brand kit — palette, type, logo rules, tone — into constraints the generator has to work within, then produces variants, batches and full campaign rollouts against them. The interesting problem was accountability rather than generation: a brand team can only adopt this if every asset can be traced to the prompt, model version and licensed inputs behind it, and if nothing reaches the public without a named human sign-off.
The challenge
Generative tools are fast and indifferent to brand rules, which makes them risky for the teams who most need the speed. Output drifts off-palette, ignores the space a headline needs, or quietly reuses references nobody has licensed. Teams end up checking every asset by hand, which erases the time saved, and they have no defensible answer when asked where a particular image came from.
Our solution
We made the brand kit the governing input rather than a style suggestion. Palette, type pairing, logo clear space and composition rules are enforced at generation time, and a guardrail panel reports which rules the current output satisfies and which it does not, in plain language. Every asset stores its prompt, model version and source references, so provenance is a property of the file rather than something reconstructed later. Nothing leaves the system without passing an approval queue where a named person signs off, and batch generation inherits the same kit so a rollout stays consistent across every size.
What we built
- 01Brand-constrained generationPalette, type and composition rules applied as constraints, not suggestions.
- 02Guardrail reportingPlain-language checks showing which brand rules an asset meets.
- 03Variant explorationBatches of options from one prompt, refined by describing a change.
- 04Asset provenancePrompt, model version and licensed inputs recorded with every file.
- 05Approval workflowNothing publishes without a named human sign-off and an audit trail.
- 06Campaign batchesBulk rollouts across sizes and channels from a single brand kit.
- 07Shared asset libraryApproved work searchable by prompt or tag, reused instead of regenerated.
- 08Brand kit versioningRules published as versions so past work stays explainable.
- 09Accessibility checksContrast flagged before an asset reaches layout, not after.
Desktop
Where the product doesits heaviest work.
The full-width interface, screen by screen.
Mobile
Rebuilt for onehand.
The same information hierarchy, rebuilt for a screen you hold in one hand.
More work
Other projects
(→)Start a project
Want something like Lumina for your business?
Tell us what you have in mind and we'll come back with a clear scope, a realistic timeline and an honest opinion on the best way to build it.
Project brief