What video format does the upscaler require?

The upscaler requires MP4. You can upload an MP4 directly or pick a clip you already generated in Workroom without downloading it first.

What resolutions can I upscale to?

You can choose 720p, 1080p, or 1440p as the target output. The AI rebuilds detail to match the resolution you select.

Does it actually add detail or just enlarge the video?

It adds detail. The model reconstructs texture and sharpens each frame, so the result looks genuinely clearer — not the same footage stretched to a larger size.

Can I upscale videos I didn't generate in Workroom?

Yes. You can upload any video — it doesn't have to come from Workroom. The upscaler works on real footage, AI-generated clips, and compressed exports alike.

Do I need to configure anything before generating?

No. The only choice is the target resolution. There are no sliders, denoise settings, or model selections — pick the resolution and click Generate.

Is the upscaler good for AI-generated video specifically?

Yes — it's tuned for it. Generative video often comes out soft and low-res, and the upscaler is built to recover the crispness and detail those models tend to lose.

AI Video Upscaler
Sharper Footage, Higher Resolution

Upload a video and get a sharper, higher-resolution version — up to 1440p. Fixes the soft, low-res look AI generation leaves behind, with no settings to tune.

Try in Workroom

Sharper, cleaner, ready to publish

Brings back lost detail

AI-generated and compressed video often comes out soft — blurred edges, mushy texture, washed-out faces. Upscale reconstructs the fine detail instead of just stretching pixels, so the footage looks captured, not rendered.

Higher resolution on demand

Choose your target — 720p, 1080p, or 1440p — and the AI rebuilds the frame at that resolution. Small previews become clips that hold up full-screen, on social, or on a big display.

One step inside your workflow

No separate app, no export and re-import. Generate a clip in Workroom, upscale it right there, and keep moving to lipsync, edit, or publish without leaving the pipeline.

Key Features

Frame-by-frame detail recovery, target resolution control, one-click processing

Frame-by-frame detail recovery

The model enhances every frame individually — sharpening edges, restoring texture, and cleaning up compression artifacts. The result stays consistent across the whole clip, not just on still frames.

Try it free

Pick your output resolution

Set the target to 720p, 1080p, or 1440p depending on where the video is going. The AI rebuilds detail to match the resolution you choose — no manual scaling, no guesswork.

Try it free

Built for AI-generated video

Generative video tends to look soft and low-res straight out of the model. Upscale is tuned to fix exactly that — recovering crispness and detail so your generated clips look finished.

Try it free