A small image printed across a wall looks soft, because there was never enough detail in it to fill that much paper. AI enhancement reconstructs plausible detail so the print holds up at size.
Enhance your image with an offline AI model for a sharper, higher-quality export. ✨
It runs on your device
The model runs locally. Your image is not uploaded anywhere — not to Docuslice, not to a third party. Speed depends on your hardware rather than on a queue somewhere.
How the model gets there differs by platform:
- Android and iOS — the models ship inside the app. Nothing is downloaded, and enhancement works with no connection at all.
- Browser and desktop — the model is fetched once and cached for later use, alongside the runtime that executes it. The first run therefore needs a connection; later runs use the cache.
Using it
- Select the image on the canvas.
- Choose Enhance in the image toolbar.
- Choose a model (below) and start.
- Wait. Progress is shown, and you can cancel — cancelling leaves the original untouched.
The app may prompt you before you ever open the tool. If an image is too small for the size you are printing at, it flags it:
Low Resolution Image — This image may appear blurry when printed. Enhance it to improve the print quality.
That warning is based on your actual poster size, so it appears when it is relevant and not otherwise.
The two models
| Model | Trade-off |
|---|---|
| Mini (Fast) | Quicker and lighter on memory. Good for a first look, and for large images on modest hardware. |
| Max (High Quality) | Better reconstruction, slower, and needs more memory. Use it for the version you are actually printing. |
Start with Max if your device can manage it. Fall back to Mini if it is slow or runs out of memory.
Before you start
AI can make mistakes. Results may vary and processing may take a moment. For best results, close other apps to free up memory before enhancing.
That advice is worth following on phones especially. Enhancement is memory-hungry, and the app caps the output to protect against running out — very large images are enhanced up to a limit rather than without bound.
If your image is already large, the app says so:
Image is already high resolution. Enhancement may not be needed.
It will still run if you ask it to. It will rarely help.
Reverting
Reopen Enhance on the image and, with the model you used still selected, the button reads Revert to Original. Press it to restore the image you imported. Pick the other model instead and it offers a fresh enhancement rather than a revert.
On Android and iOS this survives closing and reopening the project. In the browser and desktop apps it lasts for the session, like everything else there — see projects and where your work is saved.
Use it if the enhancement invented something you do not like — over-smoothed skin, plastic-looking texture, mangled small text. That is the failure mode to watch for, and it is why the app asks you to look rather than applying enhancement automatically.
What it cannot do
- It cannot recover information that was never captured. It produces a plausible reconstruction, not the photo you would have taken at higher resolution.
- It is not a substitute for a better source file. If the original exists at full size, use that.
- It does not help vector artwork. SVGs have no Enhance option at all. Imported PDF pages do, but they are already rendered at full print resolution, so it rarely changes anything — export them as vector instead.