MiniPx vs VanceAI — Free, Fast & Private
VanceAI sells credits; MiniPx sells a month of runs on top of a free layer that has no meter at all. VanceAI is a cloud AI image toolkit — enlarging, denoising, sharpening, restoring, cutting out — reachable on the web, through a Windows desktop application and through an API. Everything it does is an AI operation, so everything it does is billable, and the mental arithmetic before you press a button is "how many credits will this cost me". MiniPx splits the product in two: a browser layer that costs nothing and is never counted, and a Pro cloud layer that is.
What is free, and how free
The unmetered half of MiniPx is not a trial. Compression, format conversion, resizing, cropping, PDF work, EXIF stripping and roughly two hundred other browser tools run entirely on your own device, for as long as you like, with no account and no counter — because they cost us nothing to run. The image converter and the image compressor are the same tools for a free visitor and a subscriber. That is a structural difference, not a generous free tier: a browser cannot bill you for its own CPU. Free batches are limited to five files at a time, and the three on-device AI tools are metered separately with a small free allowance.
What a Cloud HD run is, exactly
One photo through one cloud tool is one run. Pro is $3.24/month or $29/year on the launch promo ($4.99 and $39 at list, excluding tax) and carries 150 Cloud HD runs a month on Monthly and 200 on Annual. Of those, 10 may be HD runs — object removal, the Cloud HD four-times upscale, and a restoration with the colorize step chained onto it. The allowance resets on the first of the month in UTC and does not roll over, there is a ceiling of thirty runs an hour, and a run that fails is not charged. Nothing is deducted for a preview, and nothing is deducted twice for one photo.
Mapping the jobs onto MiniPx tools
Five cloud jobs cover what most people buy credits for. An old, soft portrait goes to photo restoration, which runs RestoreFormer++ on faces at the input's own size. A grainy modern snapshot goes to the AI photo enhancer, the same pipeline with the face requirement lifted. A black-and-white print goes to the photo colorizer, which runs DDColor and returns plausible — not verified — colour as a PNG at the same size. Something unwanted in the frame goes to the object remover, where you paint a mask and PowerPaint fills the scene back in. And a small image that needs to be big goes to the image upscaler, free on your device or HAT-L four-times on our GPU for Pro.
Where VanceAI is the stronger buy
Volume, automation and a desktop binary are all on VanceAI's side. It has a public API, so it can sit inside a pipeline that MiniPx cannot reach; it has a Windows desktop application for people who would rather not work in a browser; and its catalogue of AI operations is broader than our five cloud tools. If you are processing a catalogue of hundreds of product shots on a schedule, a credit pack you can top up beats a fixed monthly allowance with an hourly ceiling. MiniPx is the better fit for a person with a folder of photographs, not a queue of them.
Where the file goes in each case
Cloud HD is the only part of MiniPx that uploads, and it is off until you switch it on. A Cloud HD photo goes to our server for that one run, is processed in the United States on Cloudflare R2 storage and Modal GPUs, travels encrypted, and is deleted when the run finishes, and in all cases within 24 hours. It is never used to train anything, and the interface names the model that ran. VanceAI is a cloud service, so its tools upload by nature; for what it retains and for how long, read their privacy policy — we do not characterise it here. If a photograph is sensitive enough that you would rather it stayed put, use the on-device half of MiniPx, which is most of it.
Compare also
Related head-to-heads: MiniPx vs Let's Enhance, MiniPx vs Hotpot.ai, MiniPx vs Remini.