Free Image Upscaler
Image Upscaler enlarges photos, drawings and screenshots 2×, 3× or 4× with Real-ESRGAN AI models, up to 16,384 pixels per side. Pick Photo, Anime & art or, with WebGPU, Max detail, preview an area to check the result and time, then download PNG, JPG or WebP. Several images can be queued and saved as one ZIP.
- Scale:
- 2×, 3× or 4×
- Max output:
- 16,384 px per side
- Models:
- Photo, Anime & art, Max detail (WebGPU)
- Price:
- Free
- Runs:
- AI model in your browser
- Uploads:
- None: files never leave your device
- Account:
- Not needed
Quick answers
Is my photo uploaded to a server?
No. The AI model is downloaded once from Hugging Face and runs inside your browser; your pictures never leave your device.
How large can the result be?
Up to 16,384 pixels per side. Depending on the memory of your device the limit is 16, 32 or 64 megapixels; the tool tells you before starting if a picture is too big for the scale you chose.
Processed on your device: your files are never uploaded. How it works
Network activity in this tab
What this page has requested since you opened the site in this tab. An upload is any request that sends data: POST, PUT, PATCH or a request with a body.
No uploads: nothing has been sent from this page.
No network requests yet.
To see every request, open your browser's developer tools (Network tab).
How to use the Image Upscaler
- Drop one or more images (photos, drawings, screenshots).
- Pick the model (Photo or Anime & art) and how much to enlarge: 2×, 3× or 4×.
- Click the picture to choose an area and press Preview area to check the result and the time it will take.
- Press Upscale, then compare with the slider and download the result (PNG, JPG or WebP).
Credits & open source
Open-source AI model running in your browser. Thank you to the people who make these projects.
- ONNX Runtime WebExecutes ONNX models with WebGPU or WebAssembly (self-hosted from /vendor/ort). © Microsoft Corporation · MIT
- fflateFast ZIP compression for batch downloads. © Arjun Barrett · MIT
Web platform APIs used: Web Workers, WebAssembly, WebGPU, Canvas 2D, createImageBitmap, File API.
- Real-ESRGAN general x4 v3 (ONNX) by Xintao Wang et al. (weights); ONNX export by SkillSafe (BSD-3-Clause) — AI model for photos (default)
- Real-ESRGAN anime video v3 (ONNX) by Xintao Wang et al. (weights); ONNX export by SkillSafe (BSD-3-Clause) — AI model for anime, drawings and screenshots
- Real-ESRGAN x4plus (ONNX) by Xintao Wang et al. (weights); ONNX export by SkillSafe (BSD-3-Clause) — AI model for maximum detail (WebGPU only)
Full license texts: third-party notices · all credits
Frequently asked questions
Why is it slow on my computer?
Without WebGPU the model runs on the processor, which is much slower. Recent Chrome and Edge use the graphics card through WebGPU and are often 20 times faster. The preview tells you how long the full picture will take.
Which model should I choose?
Photo works for most pictures and also cleans noise and JPEG blocks. Anime & art keeps lines sharp and colors flat on drawings, logos and screenshots. Max detail (WebGPU only) adds more texture to photos but is about ten times slower.
Does upscaling recover real details?
No AI can recover information that is not in the picture: the model draws plausible edges and textures. Text and faces that are only a few pixels tall may come out wrong.