21 · AI integration · Desktop app
RENEKI: AI image generation for teams, as a desktop app.
RENEKI is a vona desktop app for macOS and Windows, built for LANAKILA. Teams use it to create new image motifs from product photos – with one-click prompt building blocks, tokens and costs per image, and a web app that centrally manages users, models and building blocks.
A vona product – built for LANAKILA, the sports brand of vona owner Sylvia Michalk · project from 2026
- Desktop app
- macOS & Windows
- Electron
- React
- SQLite
- PHP backend
- Nano Banana (Gemini)
- macOS + Windows
- one desktop app, local and usable without a server connection
- Up to 8
- reference images per generation plus a main image – also straight from the shop
- Since 06/2026
- in productive use
01 · Starting point
New product images need motifs, variants and formats – fast and in the brand's style. Online AI image services are flexible, but hard to fit into a team workflow: there is no access control, no cost overview, no shared prompt base and no direct route into the shop system.
02 · Why desktop instead of browser
The design process happens on the computer – that is where the templates are: for example the design of a triathlon suit from Affinity Designer or Illustrator, cutting patterns and reference images. A web version would have meant uploading all of that first. RENEKI therefore works locally: more secure, because no design or cutting pattern sits online, and faster, without uploads and without timeouts. Only the result is uploaded – the generated product images, which can be assigned directly to a product in the shop.
Teamwork still works: projects with all their image data can be shared between several RENEKI installations. One designer works on a Mac, another in a different city on a PC – shared projects can be imported and checked out on any installation. The middleware manages, among other things, the logins of the installations.
03 · Solution
The desktop app for macOS and Windows: every piece of work belongs to a project with gallery and history. The main image and up to eight references arrive by drag and drop, from the shop or from an image already generated. The prompt builder combines camera, light, style, background and atmosphere with one click, plus a prompt library with templates. Each image shows model, tokens and costs; results can be emailed, exported, assigned to an existing WooCommerce product – or RENEKI clones a product in the shop and assigns the new images to the clone.
The typical case at LANAKILA: a cycling jacket is already in the shop with photos, and a new design for exactly this cut is launched at the same price – but without a product photo yet. The main image is the flat new design, exported from Affinity as PNG or JPG. The reference is an existing photo of the jacket in the old design: with Nano Banana and the predefined prompts, the AI dresses the model in the new design. Further references can set the style or the scene. If the image works, the existing product is cloned with one click and gets the new images.
The second use case – social media: existing product photos are turned into visuals for Instagram and Facebook. The references come straight from the product images in the shop or by drag and drop from the computer; the prompt describes the target image. One example: a triathlon suit for women and men has two plain product photos – the target image shows both people in the LANAKILA design on road bikes in the Alps. Format (1:1, 5:4, 4:5, 9:16, 16:9), resolution (1K, 2K) and creativity are adjustable, so the same material becomes a post or a reel.
Teamwork: a project is shared with one click as an immutable snapshot; colleagues check out their own copy – later changes to the original never overwrite it.
The web app centrally manages users and rights, providers and models, prompt building blocks, releases and shops. A dashboard shows generations, errors, token usage and costs; every generation can be traced with user, model and cost. Messages to the team appear in the app as system notifications, with read status.
04 · Result
One tool from idea to finished product image – as a team, with transparent costs and central management. The app works local-first: images, history and prompts are stored locally in SQLite, API keys and mail credentials in the operating system's keychain and are never sent to the server. The backend calculates the cost of each generation from the stored model prices. New providers can be added through a uniform interface. In productive use since 06/2026.
05 · Workflow
- Web app & backend: Users, rights, providers, models, Prompt blocks, releases, shops, Tracking and costs on the server, API keys stay local
- FROM PRODUCT PHOTO TO AI IMAGE: Project (main image + up to 8 references) → Prompt builder (camera, light, style combined with a click) → Model (Nano Banana & co. via Google AI Studio) → Gallery (every image with tokens and costs) → Next (email, export or shop product)
- TEAMWORK: Share project (snapshot with images) → Check out (own copy, nothing is overwritten)
06 · How RENEKI works
Schematic, animated diagrams of the workflow with sample data. The screenshots below show the real app.
Workflow: source file, prompt, generate, review and result – plus sharing in the team via the middleware and the web app behind the scenes Second use case – social media: existing product photos from the shop become a photo composite from a description, as a post or reel in 1:1, 5:4, 4:5, 9:16 or 16:9 Why a desktop app: designs, cutting patterns and references stay on the computer, only the finished image goes to the shop
07 · Insights
Related services
We can do this for you, too.
The services behind this project — with examples, process and frequently asked questions.
AI integration
We implement artificial intelligence where it creates real value in everyday work — from intelligent chatbots and AI workflows to the automation of recurring office tasks.
Learn more →App development
Native and cross-platform apps for iOS and Android — and desktop apps for macOS and Windows. Thought through in UX and design, strong in performance.
Learn more →Keep reading
Background from practice.
Blog
Prompt Engineering: Eight Principles, a Template and Examples
Eight principles for better prompts with a before-and-after example, a template to take away, notes on prompt injection and a prompt library for your team.
Read more →Blog
Corporate Design in the Age of AI
If AI tools can generate logos — what role is left for strategic corporate design? Putting it in perspective.
Read more →Next project
Your project here?
Tell us about your problem. The first conversation is of course free of charge and without obligation.
Request a project →