# Free AI Image Upscaler > Kenerate AI's image upscaler is a free-to-start online tool that makes a picture bigger and sharper with AI. It has seven engines: Real-ESRGAN (the fast default), Image Upscaler for 2K, 4K or 8K, Kenerate Upscale for an exact size up to 128 MP, Clarity Pro for creative detail, Phota Enhance for old and blurry photos, plus Ultimate and Recraft. New accounts get 25 free credits and there is no watermark. URL: https://kenerateai.com/ai-image-upscaler Publisher: Kenerate AI (https://kenerateai.com) Last updated: 26 September 2026 ## Key facts - Speed: About 15–80 seconds per image - Free to start: 25 credits on sign-up, no watermark - Max size: 8K, or up to 128 MP - Engines: 7 — faithful, creative and restore ## Steps 1. Add an image — Upload a JPG, PNG or WEBP up to 25 MB — a photo, a product shot, a screenshot or AI art — or try one of the samples. 2. Pick an engine and a size — Real-ESRGAN is the fast default. Choose Image Upscaler for 2K, 4K or 8K, Kenerate Upscale for an exact size up to 128 MP, Clarity Pro to add new detail, or Phota Enhance to repair old and blurry photos. 3. Compare and download — Drag the slider to check the result against the original, then download it — no watermark. Send it on to Edit Studio or the background remover if you need to. ## In depth ### Why a normal resize makes pictures blurry Classic resizing — nearest-neighbour, bilinear, bicubic (Keys, 1981) and Lanczos (Duchon, 1979) — creates each new pixel as a weighted average of the pixels around it [1][2]. No new fine detail is created, so when you enlarge, edges are simply spread across more pixels and the image looks soft. Nearest-neighbour keeps hard edges but turns curves into visible blocks; Lanczos is sharper than bicubic but can add halos (ringing) along edges [3]. There is also a hard limit: a sampled image only holds detail up to half its sampling rate (Shannon's sampling theorem) [4]. Many different sharp images shrink to the same small one, so any upscaler has to choose a plausible answer rather than recover the true one. ### How AI super-resolution learned to add detail SRCNN (Dong, Loy, He and Tang, ECCV 2014) is widely cited as the first deep-learning super-resolution network: one small convolutional network learns the whole mapping from low- to high-resolution [5]. SRGAN (Ledig et al., CVPR 2017) added a perceptual loss and an adversarial (GAN) loss and produced the first photo-realistic 4× results [6]; ESRGAN (Wang et al., 2018) improved it and won the PIRM2018 super-resolution challenge — while noting that GAN-hallucinated details can come with artefacts [7]. Real-ESRGAN (Wang, Xie, Dong and Shan, ICCV Workshops 2021) — the default engine in Kenerate's upscaler — was trained on synthetic “high-order” degradations: blur, noise, resizing and JPEG compression applied, then applied again, so it copes with messy real-world photos instead of clean lab images [8]. It is open source under the BSD-3-Clause licence and has a smaller model tuned for anime [9]. ### Faithful vs creative: the perception–distortion trade-off Blau and Michaeli (CVPR 2018) proved that no restoration method can be both maximally faithful to the true pixels and maximally natural-looking — improving one eventually costs the other [10]. That is why Kenerate offers both kinds of engine. Faithful engines (Real-ESRGAN, Image Upscaler, Kenerate Upscale, Ultimate) stay close to the input. Creative engines are closer to diffusion upscalers such as Stability AI's SD x4 upscaler [11], StableSR [12] and SUPIR [13], which regenerate detail from a learned image prior; StableSR even exposes a single dial to balance quality and fidelity — the same idea as Clarity Pro's creativity slider. ### Why upscaled faces and text need a second look In June 2020 the PULSE face “depixelizer” (Menon et al., CVPR 2020) turned a pixelated photo of Barack Obama into a white man, because it searched a face generator for any face that matched the pixels — not the real person [14][15]. In 2024 a Washington State court excluded AI-“enhanced” video from a criminal trial, finding it showed what the model “thought should be shown” [16]. The lesson for everyday use: AI upscaling is excellent for making photos, products and artwork look better, but it is not evidence. Keep faithful engines for anything factual, compare faces and text at 100%, and keep your original file. References: [1] Keys (1981) — Cubic convolution interpolation, IEEE Trans. ASSP: https://doi.org/10.1109/TASSP.1981.1163711 [2] Duchon (1979) — Lanczos filtering in one and two dimensions: https://journals.ametsoc.org/view/journals/apme/18/8/1520-0450_1979_018_1016_lfioat_2_0_co_2.xml [3] Image scaling — Wikipedia: https://en.wikipedia.org/wiki/Image_scaling [4] Shannon (1949) — Communication in the Presence of Noise: https://www.math.unl.edu/~bdeng1/Teaching/math943/Topics/Genetics/Shannon1949.pdf [5] Dong et al. — SRCNN (ECCV 2014 / TPAMI 2016): https://arxiv.org/abs/1501.00092 [6] Ledig et al. — SRGAN (CVPR 2017): https://arxiv.org/abs/1609.04802 [7] Wang et al. — ESRGAN (ECCV Workshops 2018): https://arxiv.org/abs/1809.00219 [8] Wang et al. — Real-ESRGAN (ICCV Workshops 2021): https://arxiv.org/abs/2107.10833 [9] Real-ESRGAN — GitHub (BSD-3-Clause, anime model): https://github.com/xinntao/Real-ESRGAN [10] Blau & Michaeli — The Perception-Distortion Tradeoff (CVPR 2018): https://arxiv.org/abs/1711.06077 [11] Stability AI — Stable Diffusion x4 upscaler: https://huggingface.co/stabilityai/stable-diffusion-x4-upscaler [12] Wang et al. — StableSR (IJCV 2024): https://arxiv.org/abs/2305.07015 [13] Yu et al. — SUPIR (CVPR 2024): https://arxiv.org/abs/2401.13627 [14] Menon et al. — PULSE (CVPR 2020): https://arxiv.org/abs/2003.03808 [15] The Verge — The Obama depixelizer and AI bias: https://www.theverge.com/21298762/face-depixelizer-ai-machine-learning-tool-pulse-stylegan-obama-bias [16] NBC News — Washington judge blocks AI-enhanced video evidence (2024): https://www.nbcnews.com/news/us-news/washington-state-judge-blocks-use-ai-enhanced-video-evidence-rcna141932 ## FAQ Q: Is the Kenerate AI image upscaler free? A: It is free to start: new accounts get 25 free credits on sign-up to test the engines on your own images. Bigger jobs use one-time credit packs that never expire — no subscription, no watermark. Q: What is the best AI image upscaler? A: It depends on the picture. For everyday photos, screenshots and AI renders a faithful engine such as Real-ESRGAN or Image Upscaler is best because it keeps the image as it is. For faces and campaign visuals where you want extra fine detail, a creative upscaler like Clarity Pro is better. For old, blurry or damaged photos use a restoration model like Phota Enhance. Kenerate has all three in one tool, so you can try more than one on the same image. Q: How big can I upscale an image? A: Image Upscaler and Ultimate output 2K, 4K or 8K. Kenerate Upscale lets you pick an exact size from 4 to 128 megapixels, and Clarity Pro goes up to 64 megapixels. Q: Does upscaling add real detail? A: A faithful upscaler rebuilds edges and textures that are implied by the pixels you already have, so the result looks sharper than a normal resize. It cannot recover detail that was never captured. Creative engines (Clarity Pro, Recraft) can invent new fine detail — which looks great but may change small things like skin texture or tiny text, so check the result. Q: Which upscaler should I use for faces? A: Start with Image Upscaler or Real-ESRGAN, which keep the face as it is. If the photo is old or out of focus, use Phota Enhance. Clarity Pro adds the most detail but, with the creativity dial high, it can change features — keep it near 0 for portraits. Q: Can I upscale anime and illustrations? A: Yes. Real-ESRGAN keeps clean line art and flat colours, which is what you want for anime, manga, logos and cartoons. Q: What file types can I upload and download? A: Upload JPG, PNG or WEBP up to 25 MB. Most engines let you download PNG, JPG or WEBP. Q: Is there a watermark? A: No. Results from free and paid credits are the same clean file. Q: Can I use the upscaled images commercially? A: Yes, as long as you have the rights to the original image. You own your results. Q: How long does it take? A: About 15 to 80 seconds depending on the engine: Kenerate Upscale is the fastest (~14 s), Real-ESRGAN about 25 s, Clarity Pro about 80 s. ## Related pages - Upscale image: https://kenerateai.com/upscale-image - AI image enhancer: https://kenerateai.com/ai-image-enhancer - Upscale image to 4K: https://kenerateai.com/upscale-image-to-4k - Increase image resolution: https://kenerateai.com/increase-image-resolution - HD photo converter: https://kenerateai.com/hd-photo-converter - Unblur image: https://kenerateai.com/unblur-image - Sharpen image: https://kenerateai.com/sharpen-image - Unpixelate image: https://kenerateai.com/unpixelate-image - Photo restoration: https://kenerateai.com/photo-restoration - Uncensored image enhancer: https://kenerateai.com/enhance-ai - Enhance AI art: https://kenerateai.com/enhance-ai-art - Anime upscaler: https://kenerateai.com/anime-upscaler - Bulk image upscaler: https://kenerateai.com/bulk-image-upscaler - Remove glare from photo: https://kenerateai.com/remove-glare-from-photo - Restore old photos: https://kenerateai.com/how-to-restore-old-photos - Best photo restoration software: https://kenerateai.com/best-photo-restoration-software - Denoise photos: https://kenerateai.com/how-to-denoise-photos - Increase image resolution: https://kenerateai.com/how-to-increase-resolution-of-an-image - Sharpen an image in Photoshop: https://kenerateai.com/how-to-sharpen-an-image-in-photoshop - All upscaler alternatives: https://kenerateai.com/image-upscaler-alternatives - Remini alternative: https://kenerateai.com/remini-alternative - Let's Enhance alternative: https://kenerateai.com/lets-enhance-alternative - upscale.media alternative: https://kenerateai.com/upscale-media-alternative - Upscayl online: https://kenerateai.com/upscayl-online - Bigjpg alternative: https://kenerateai.com/bigjpg-alternative - VanceAI alternative: https://kenerateai.com/vanceai-alternative - waifu2x alternative: https://kenerateai.com/waifu2x-alternative - Magnific AI alternative: https://kenerateai.com/magnific-ai-alternative - Topaz Labs alternative: https://kenerateai.com/alternatives/topazlabs-alternative ## Sources - Real-ESRGAN (Wang et al., 2021) — GitHub: https://github.com/xinntao/Real-ESRGAN - Super-resolution imaging — Wikipedia: https://en.wikipedia.org/wiki/Super-resolution_imaging