BigSplit.
Workspace

Video · Beta

Video Upscaler

Increase video resolution with AI processing. Choose 2× or 4×, inspect the result and download an MP4.

Shared processing queue
Processor unavailable · submissions stay queued
Beta · AI video upscaling

Increase frame resolution with Real-ESRGAN. Fine detail is estimated, and moving textures can flicker. Preview the full result before using it.

For an iPhone spatial MOV, use Spatial Video to Gaussian Splat to reconstruct a 3D scene from both camera views.

Output: H.264 MP4 with the source audio. Frame timing may be regularised for variable-frame-rate sources. HDR, wide-gamut video and outputs above 200 MB are not supported in this beta.

One job runs at a time. Requests expire if they cannot start within 12 hours. Email delivery is unavailable; return to this page in the same browser and keep its site data.

How to use Video Upscaler

  1. 1Choose an SDR clip up to 30 seconds, 720p and 30 fps.
  2. 2Select 2× or 4× and submit the video to the processing queue. Output is limited to 4K.
  3. 3Watch the result for altered detail or flicker, then save the MP4 within 24 hours.

Before you start

Does upscaling restore the original detail?

No. Real-ESRGAN estimates higher-resolution detail from each frame. Faces, text and moving textures can change or flicker. Inspect the result before using it.

What are the beta limits?

Inputs up to 100 MB, 30 seconds, 720p and 30 fps. Output up to 4K and 200 MB. HDR and wide-gamut video are not supported.

Where does processing happen?

Your video is uploaded over HTTPS to BigSplit and processed in a shared private queue. It may wait while the processor is busy or offline.

Is the audio kept?

The source audio is retained. AAC is copied where possible; other formats are converted to AAC. Variable frame timing may be regularised.

Read the guides →File processing & privacy →