What is Topaz Video Upscale?
Topaz Video Upscale is the API form of Topaz Video AI's upscaling engine. Send a video_url, choose an output resolution tier from 480p up to 8K (7680 px on the long edge), and optionally pick one of 21 Topaz model presets, from Proteus for general footage to Iris for face recovery and Dione for deinterlacing. The same task can chain a frame-interpolation pass for higher frame rates or slow motion. Source FPS is preserved unless you set target_fps or turn interpolation on. Enhancement sliders such as compression recovery, detail, noise, and grain go in extra_settings; anything you leave out is set automatically by Topaz.
Key features of Topaz Video Upscale
Four features cover what you'll set on a typical job.
Six output tiers up to 8K
480p, 720p, 1080p, 1440p, 4K, and 8K, set with resolution. The long edge maps to exact pixels and the short edge follows the source aspect ratio. Use 1080p for SD archives and 4K or 8K for delivery masters.
21 Topaz model presets
Pick the model for the footage: Proteus for general use, Rhea for fine detail, Iris for faces, Gaia for CGI and animation, Dione for interlaced TV and DV, Nyx for denoise, Themis for motion deblur, and Colorize for black-and-white film.
Frame interpolation and slow motion
Set frame_interpolation to true with an Apollo or Chronos preset to raise the frame rate (15 to 240 fps) or slow the video down up to 16x in the same job. No second pass, no intermediate file.
Pipeline-friendly task shape
Same POST /v1/tasks shape as every other media model in the API. Submit a job, get a task_id, poll or wait for a webhook. Easy to drop into an existing batch processor.
What Topaz Video Upscale is best for
Archive remasters
Push SD or HD masters to 1080p, 4K, or 8K delivery files in a single call, with Proteus or Rhea doing the reconstruction.
Legacy TV and DV footage
Dione presets deinterlace broadcast and DV sources while upscaling, so old tapes come out progressive and sharper in one job.
Frame rate conversion and slow motion
Interpolate 24 fps source to 60 fps for sports or gaming edits, or slow an action shot down 2x to 16x without a second tool.
Face recovery
Iris presets restore faces in compressed or low-quality interviews and user-generated video.
CGI and AI-generated video
Gaia presets keep clean edges and flat colour regions, which suits animation and the output of video generation models.
Batch normalization
Drive the same resolution tier and preset across an asset library so the whole catalogue lands at a consistent delivery resolution.
Topaz Video Upscale use cases
Push an SD or HD archive to 1080p or 4K with one call per video, then drop the results back into the catalogue. Deinterlace a box of digitised TV recordings with a Dione preset while upscaling them to 1080p. Turn a 24 fps action shot into 2x slow motion at 4K by combining resolution: "4k" with a Chronos interpolation pass. Upscale the output of a video generation model to 4K or 8K before delivery. Stand up an upscale step inside a content tool where users upload older videos and get a higher-resolution master back.
Topaz Video Upscale limitations
Output keeps the source aspect ratio: if you need a different aspect, reframe in your NLE before or after the call. Uploads through the playground are limited to 500 MB and 300 seconds per video. Higher tiers and 50/60 fps output cost more per second, so test on a short clip before running a full archive at 8K. Interpolation settings only apply when frame_interpolation is on, and an interpolation_preset is required then.
Topaz Video Upscale API examples
Call Topaz Video Upscale from any language by POSTing to /v1/tasks. Full parameter docs live at docs.unifically.com/models/video/topaz-labs/video-upscale.
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "topaz-labs/video-upscale",
"input": {
"video_url": "https://example.com/video.mp4",
"preset": "prob-4",
"resolution": "4k"
}
}'
Successful submission returns a task_id. Poll GET /v1/tasks/<task_id> or set a callback_url on the request to receive the finished video URL.
FAQs
People also ask about Topaz Video Upscale
Topaz Video Upscale is the API form of Topaz Video AI's upscaling engine. Send a video_url, pick an output resolution tier (480p up to 8K) and optionally a Topaz model preset, and the model returns a higher-resolution video. One endpoint covers both upscale-only jobs and upscale plus frame interpolation.
Leave target_fps unset and do not turn on frame_interpolation. The output then keeps the source video's frame rate.
Six tiers set with resolution. 480p (854 px long edge), 720p (1280 px), 1080p (1920 px), 1440p (2560 px), 4k (3840 px), and 8k (7680 px). The short edge is derived from the source aspect ratio, so the output keeps the source orientation.
The preset field takes 21 Topaz model codes. Proteus (prob-4, the default) for general footage, Rhea for high-fidelity detail, Theia for sharpening, Iris for face recovery, Gaia for CGI and animation, Artemis for denoise plus sharpening, Dione for deinterlacing, Nyx for denoise, Artemis AA for moire, Colorize for black-and-white footage, and Themis for motion deblur.
Yes. Set target_fps for a plain frame-rate change, or set frame_interpolation to true with an interpolation_preset (Apollo or Chronos) to chain an interpolation pass after the upscale. Inside interpolation_settings, slowmo (1 to 16) slows the video down and fps (15 to 240) raises the frame rate without changing duration.
Same engine family. The API exposes the Topaz model presets, the enhancement sliders (through extra_settings), and frame interpolation in a single task. Use the API for batch automation and pipelines; use desktop when an editor wants to scrub and hand-tune every shot.
Per finished output second on topaz-labs/video-upscale, tiered by output resolution, with 50/60 fps output billed at a higher tier. See the pricing page for the current rates.
