Free Video Upscaler
Video Upscaler enlarges short videos 2× or 4× with Real-ESRGAN AI models for real footage or anime, with output up to 4K. Clips can be up to 60 seconds and 720p, a Fast mode without AI is included, and the sound is kept. A one-frame preview compares the result with a plain browser resize and estimates the total time.
- Input:
- Up to 60 s and 720p
- Output:
- 2× or 4×, up to 3840 × 2160
- Methods:
- AI · Real, AI · Anime or Fast
- Price:
- Free
- Runs:
- AI model in your browser
- Uploads:
- None: files never leave your device
- Account:
- Not needed
Quick answers
Is my video uploaded?
No. The AI model is downloaded once from Hugging Face and runs in your browser; the video never leaves your device.
How long does it take?
AI upscaling runs the model on every frame: with WebGPU (recent Chrome or Edge) expect from a fraction of a second to a few seconds per frame depending on the resolution and your graphics card, so a 10-second clip can take minutes. The preview tells you the time before you start. The Fast method takes seconds.
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 Video Upscaler
- Drop a short video (up to 60 seconds and 720p).
- Pick the method (AI for real footage, AI for anime, or Fast without AI) and 2× or 4×.
- Move the slider to a moment and press Preview frame to compare the browser resize with the result and see how long it will take.
- Press Upscale: every frame is processed on your device with a progress bar, then play and download the MP4 with the original sound.
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
- MediabunnyReads, writes and converts media files with WebCodecs (MP4, WebM, MKV, MOV, MP3, WAV, Ogg...). © Vanilagy and contributors · MPL-2.0
Web platform APIs used: Web Workers, WebAssembly, WebGPU, WebGL, WebCodecs, OffscreenCanvas, Canvas 2D, File API.
- Real-ESRGAN general x4 v3 (ONNX) by Xintao Wang et al. (weights); ONNX export by SkillSafe (BSD-3-Clause) — AI model for real footage
- Real-ESRGAN anime video v3 (ONNX) by Xintao Wang et al. (weights); ONNX export by SkillSafe (BSD-3-Clause) — AI model for anime and cartoons
- Real-ESRGAN general x4 v3, half precision (ONNX) by Xintao Wang et al. (weights); fp16 ONNX export by FuryTMP (BSD-3-Clause) — faster AI model for real footage on WebGPU with fp16
- Real-ESRGAN anime video v3, half precision (ONNX) by Xintao Wang et al. (weights); fp16 ONNX export by FuryTMP (BSD-3-Clause) — faster AI model for anime on WebGPU with fp16
Full license texts: third-party notices · all credits
Frequently asked questions
Why the 60-second and 720p limits?
Every frame goes through the AI model and the result is kept in memory until you download it. The limits keep the processing time reasonable and the output at most 4K, which browsers can encode.
Is the sound kept?
Yes. The audio track is copied without re-encoding when the output format accepts it, otherwise it is converted to AAC or Opus.
Does AI recover lost detail?
No model can recover information that is not in the video: it draws plausible edges and textures and removes compression noise. Tiny text and distant faces may come out wrong.