Real-ESRGAN vs Upscale.media
Real-ESRGAN
github.comUpscale.media
upscale.mediaSide by side
| Figure | Real-ESRGAN | Upscale.media |
|---|---|---|
| Price from | — | $5.83/mo |
| Free plan | — | Yes |
| Free trial | — | — |
| Open source | Yes | — |
| Web | Yes | Yes |
| Windows | Yes | — |
| macOS | Yes | — |
| Linux | Yes | — |
| API | — | Yes |
| Self-hosted | Yes | — |
| Plans published | 0 | 12 |
| Spec score · Image Upscaling Software | 8.2 (#33) | 9.1 (#3) |
| Spec score · AI Image Upscalers | 7.6 (#24) | 8.4 (#12) |
| Batch processing | Yes | Yes |
| Desktop app | Yes | No |
| Free plan | Yes | Yes |
| Input formats | The executable accepts JPG, PNG, and WebP images. | Supports MP4, MOV and WebM video formats. |
| Integrations | The README says Real-ESRGAN was integrated into Hugging Face Spaces with Gradio and links to a web demo. | The page states that APIs integrate with existing systems. |
| Maximum scale | custom | 8x |
| Upscaling method | ai | ai |
Plans
Upscale.media
- Free PlanFree
- Subscription Plan$69.99 a year
- 60 Credits$5.99 a month
- 125 Credits$9.99 a month
- 500 Credits$24.99 a month
- 1200 Credits$54.99 a month
- Enterprise PlanPriced on request
- Free PlanFree
- One time payment$11.99, paid once
- Subscription Plan$69.99 a year
In their own words
Real-ESRGAN
Real-ESRGAN is a free, open-source project developing algorithms for image and video restoration. It offers pretrained models for general images, anime images and animation video, including AnimeVideo-v3 and smaller anime models. Image inputs include JPG, PNG and WEBP; inference can handle alpha-channel, grayscale and 16-bit images, and outputs can be JPG, PNG or WebP. Portable inference supports scale ratios of 2, 3 or 4, with 4 as the default; the Python implementation also offers arbitrary output scaling. Portable NCNN executables are available for Windows, Linux and macOS, with required binaries and models included and no CUDA or PyTorch environment needed. These executables do not include every Python feature, including outscale. Python use requires Python 3.7 or newer and PyTorch 1.7 or newer. Online inference is available through an ARC demo and Colab demos. GFPGAN is integrated for face enhancement, and the training code supports fine-tuning on user data or paired data. The project uses synthetic training data and is released under the BSD-3-Clause license.
Upscale.media
Upscale.media enlarges and enhances images using AI, with scaling up to 8x and batch processing for multiple images. Its tools include sharpening, retouching, face enhancement, deblurring, dehazing, and restoration. It accepts PNG, JPEG, JPG, WebP, and HEIC files, and images can be processed in a browser without installing software. API access supports integration with existing systems; the page also describes a console and dashboard for administration and bulk transformations. A free Google Play app is available, but there is no desktop app. The free plan includes 3 credits and 3 downloads each month, while guests can upscale one image per day. Paid options include 60 credits for $5.99 per month and a subscription for $69.99 per year. The annual subscription renews yearly and can be cancelled anytime. Monthly credits expire at each billing-cycle end; purchased credits remain valid for one year. Logged-in images can be up to 20,000 x 20,000 at 1x, with lower limits at higher scales.