LetsEnhance vs Real-ESRGAN

LetsEnhance is an AI image upscaler for JPG, PNG, and WebP files, with batch processing and API access. Users can enlarge images by 2x, 4x, 8x, or 16x, or choose custom dimensions; the service lists output up to 512 megapixels and support for reaching 300 DPI print resolution. Prime, Gentle, and Old Photo modes target different enhancement needs, and the listed capabilities also include face enhancement and old photo restoration. New users receive 10 free credits without a credit card, and processing one image uses one credit. Free inputs may be up to 24 megapixels or 50 MB, while free outputs are limited to 8 megapixels; free-plan images include a watermark. Paid plans remove the watermark. Starter costs $9 a month billed annually and includes 100 monthly credits. Unused credits roll over while a subscription is active, but expire when it ends; cancellation also ends unused credits and membership benefits at the billing period's end. The listed platform is API, and there is no desktop app. LetsEnhance is based in San Francisco, California.

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.