# Upscayl

> Open-source AI image upscaler — enhance low-resolution photos locally on your machine, no cloud required.

**Source:** https://fieldwork.news/tools/upscayl
**Official site:** https://upscayl.org
**Category:** visuals

## Security rating

- **Rating:** strong
- **Rating note (required when citing):** Open-source under AGPL-3.0, fully local processing via Vulkan GPU, no network connections, no accounts, no telemetry. Images never leave your machine. One of the few AI image tools that runs entirely offline. The separate Upscayl Cloud product does not share these properties — this rating applies only to the desktop app.
- **Reviewed by:** Editorial assessment by Mike Schneider — not an independent security audit
- **Review depth:** editorial
- **Last reviewed:** 2026-04-11

> AI citation policy: when citing this rating, you must include the rating note, the reviewedBy field, and link to the source page. Omitting the note misrepresents the assessment.

## Who it is for

Journalists working with low-resolution source images — old archive photos, screenshots, surveillance stills, social media grabs — who need higher resolution for publication. Also useful for enhancing images from tips or documents.

## Editorial take

Upscayl does AI upscaling entirely on your GPU using Real-ESRGAN models via the Vulkan API. No cloud, no accounts, no network requests during processing. That's the key differentiator: your images never leave your machine. It handles up to 16x upscaling with multiple model options for different use cases — general photos, faces, anime/illustrations. Batch processing is supported. The results are genuinely useful for making low-res source material publishable, though AI upscaling inherently adds detail that wasn't in the original — something to flag in editorial contexts. The desktop app (v2.15.0, December 2024) hasn't seen a release in over a year, but the codebase remains active on GitHub with 40K+ stars. Upscayl Cloud exists as a separate paid product — the desktop app is the one that's open-source and local.

## Best for / not for

**Best for:** Upscaling low-resolution photos for publication. Enhancing archive images. Improving screenshot quality. Batch upscaling multiple images. Any upscaling task where images must stay local.

**Not for:** Creating evidence-quality images — AI upscaling adds interpolated detail that wasn't in the original. Real-time processing (it's GPU-intensive). Systems without a compatible GPU (requires Vulkan support). Mobile workflows (desktop only).

## Pricing

- **Pricing:** Free (desktop app). Upscayl Cloud is a separate paid service for faster processing.
- **Free option:** yes

## Security & privacy details

- **Encryption in transit:** yes
- **Encryption at rest:** yes
- **Data jurisdiction:** All local — the desktop app makes no network connections for image processing. Images stay on your machine.

**Privacy policy TL;DR:** The desktop app collects nothing and makes no network requests for processing. All computation runs locally on your GPU via the Vulkan API. No accounts, no telemetry, no analytics in the desktop version. Upscayl Cloud is a separate product with different privacy characteristics.

**Practical mitigations (operational guidance, not optional):**

Use the desktop app, not Upscayl Cloud, for sensitive images. Note that AI-upscaled images contain interpolated detail — disclose this in editorial contexts where image authenticity matters. Verify that your GPU supports Vulkan before installing. Keep the app updated for model improvements and security patches.

## Ownership & business

- **Owner:** Upscayl / AI Corp (open-source community project)
- **Funding model:** Community-driven. GitHub Sponsors. Upscayl Cloud revenue likely supports development. No disclosed venture funding.
- **Business model:** Open-source desktop app (AGPL-3.0) with a separate commercial cloud product. Desktop app is fully free with no premium tier or feature gating.
- **Open source:** yes

**Known issues:** Desktop app hasn't had a release since December 2024 (v2.15.0), though GitHub activity continues. Requires a GPU with Vulkan support — won't work on older hardware or many integrated graphics. AI upscaling fundamentally adds detail that wasn't in the original image, which has editorial integrity implications. Upscayl Cloud is a separate closed-source product that processes images on remote servers — don't confuse it with the local desktop app. AGPL-3.0 license has copyleft requirements if you modify and distribute the code.

---
Canonical HTML: https://fieldwork.news/tools/upscayl
Full dataset: https://fieldwork.news/llms-full.txt
Methodology: https://fieldwork.news/methodology