AI Video Upscaler API : Upscale Videos to 1080p, 2K or 4K
Send a clip, choose the resolution and get a higher-resolution mp4 back with its sound. Try it right here with a cost estimate before you start, then call the same model from your product with one key and one balance.
- One video, up to 3 minutes
- 1080p, 2k or 4k
- Priced from the source video
- Same model through the API
Features
Key AI Video Upscaler API Features
Two upscaling engines
Choose provider bytedance or topaz on each task. bytedance offers 1080p, 2k and 4k with frame rate and content presets; topaz offers 1080p and 4k and needs nothing but the resolution.
Output at up to 60 fps
With bytedance, set fps to 24, 30 or 60 for the upscaled video. 60 fps gives smoother motion and is billed at double the 24 or 30 fps rate.
Presets for your footage
Tell bytedance what the clip is: common, aigc for AI-generated video, short_series for short dramas, ugc for footage shot by users, or old_film. Presets don't change the price.
Sound included
Upscaled videos keep their audio track. bytedance returns an h264 mp4 and topaz returns HEVC (h265); every result also has a light h264 preview for web pages.
Clips up to 3 minutes
Source videos can run up to 3 minutes. Shorter clips are the most reliable, especially at 4k, and a clip that fails during upscaling is refunded in full.
Priced by the source video
bytedance is priced per second of the source video; topaz per frame plus a fixed amount per task. The amount is worked out from your clip within seconds, and the final charge never exceeds it.
Overview
What Is the AI Video Upscaler API?
- Model ID
upscale-video- Input
- One video
- Resolution
- 1080p / 2k / 4k
- Video length
- Up to 180 s
- Price
- from $0.00936 per second
An AI video upscaler raises the resolution of a video, so footage that was shot, compressed or generated at a low resolution can play at 1080p, 2K or 4K. The AI Video Upscaler API does this over HTTP: send one video link and the output settings, and get back a single upscaled mp4, sound included.
On NanoAPI it's the model upscale-video of the product upscale_video, with two providers. bytedance takes a resolution of 1080p, 2k or 4k, an output fps of 24, 30 or 60, and a preset that matches the content: common, aigc, short_series, ugc or old_film. topaz takes a resolution of 1080p or 4k and nothing else, and returns HEVC video.
Pricing follows the source video. Within a few seconds of creation NanoAPI reads your clip, works out the price and reserves it: per second of video for bytedance, doubled at 60 fps, and per frame plus a fixed amount per task for topaz. The studio on this page shows an estimate before you start, and the charge after success is never more than the amount reserved.
Try it in the studio at the top of this page, then call the AI Video Upscaler API at https://api.nanoapi.ai/v1 from your own app, with the same POST /v1/tasks endpoint, API key and prepaid balance as every NanoAPI model.
Pricing
AI Video Upscaler API Pricing
bytedance is priced per second of the source video by output resolution, with 60 fps at double the rate; topaz is priced per frame plus a fixed amount per task. A failed task is refunded in full.
provider: bytedance
| Resolution | 24 fps · per second | 30 fps · per second | 60 fps · per second |
|---|---|---|---|
| 1080p | $0.00936 | $0.00936 | $0.01872 |
| 2k | $0.01872 | $0.01872 | $0.03744 |
| 4k | $0.03744 | $0.03744 | $0.07488 |
provider: topaz
| Resolution | Per frame | + per task |
|---|---|---|
| 1080p | $0.000585 | $0.0936 |
| 4k | $0.00195 | $0.0936 |
bytedance: cost = price per second × source video seconds
topaz: cost = per task + price per frame × source video frames (the frame part is rounded up)
The playground shows an estimate; the exact amount is frozen a few seconds after the task is created, once the source video has been read
How it works
How to Use the AI Video Upscaler API
- 1
Step 1: Get your API key
Sign in to NanoAPI with Google and create a key on the API Keys page of the console. It authenticates every request, so keep it on your server.
- 2
Step 2: Top up your balance
Add funds on the Top up page of the console. The balance is prepaid in USD, never expires and is shared by every NanoAPI model, image and video alike.
- 3
Step 3: Create a task
Send POST /v1/tasks with model upscale-video, video_url, provider and resolution; for bytedance, add fps and preset as well. All of them are required, and topaz rejects fps, preset or 2k with a 400 error.
- 4
Step 4: Get the video
Receive the finished task once at your webhook_url, or poll GET /v1/tasks/{task_id} every 15 seconds or more. A 10-second clip upscaled to 1080p takes about two to two and a half minutes; the video is at outputs[].url.
Why NanoAPI
Why Use the AI Video Upscaler API on NanoAPI
One key, every model
Generate a clip with Kling 4.0 or Seedance 2.5, reframe it and upscale it with the same API key and balance. Every NanoAPI model goes through POST /v1/tasks and returns the same task object.
Failures free, charges capped
If the video can't be read or is too long, the task fails within seconds with nothing charged, and a failure later on is refunded in full. A successful task is never charged more than it reserved.
Webhooks or polling
Pass a webhook_url and receive the finished task once, signed to the Standard Webhooks spec (verifying is optional), or poll one task or a batch of task ids.
Safe retries and spend limits
Send an Idempotency-Key and a resent request returns the task it already created instead of charging twice. Cap each key's hourly, daily or total spending in the console.
A console for every request
See each request with its status, input, cost and result, your daily usage by product and every movement of your balance, and manage keys and the webhook secret.
Try online, then integrate
The studio on this page runs the same model at the same prices and shows the estimated cost before you start. The docs also come as one llms.txt file for AI coding assistants.
Use cases
What You Can Build With the AI Video Upscaler API
AI-generated video in 4K
Upscale clips from text to video and image to video models with the aigc preset, so they hold up on big screens and in ads.
Short dramas and series
Use the short_series preset to bring episodes up to 1080p or 2K for streaming apps and social platforms.
User-shot footage
Raise the resolution of phone footage from creators and customers with the ugc preset before it goes into compilations, reviews or ads.
Old film and archive clips
Give archive footage a second life at a higher resolution with the old_film preset, for documentaries, anniversary edits and remasters. Your app can call the AI Video Upscaler API for each clip as it's uploaded.
FAQ
AI Video Upscaler API FAQ
Can’t find what you need? Write to us.
01What is the AI Video Upscaler API?
An HTTP API that upscales a video to 1080p, 2K or 4K with AI. On NanoAPI you send POST /v1/tasks with model upscale-video, a video_url, a provider and a resolution (plus fps and preset for bytedance), and get one upscaled mp4 back by webhook or by polling.
02How much does video upscaling cost?
It depends on the source video and the provider. bytedance charges per second of the source video by output resolution, and 60 fps costs double; topaz charges per frame plus a fixed amount per task. The price is reserved a few seconds after creation and the final charge is never higher. Live prices are in the table above, on the pricing page and in GET /v1/products.
03bytedance or topaz: which should I use?
bytedance offers 2k as well as 1080p and 4k, lets you set the output fps and a preset for the content, and returns h264. topaz offers 1080p and 4k, without fps or preset, and returns HEVC (h265). Both keep the audio, and both are priced from the source video.
04How long can the video be?
Up to 3 minutes; a longer video fails within seconds with video_too_long and nothing is charged. Within that limit shorter clips are the most reliable, especially at 4k, and a clip that fails during upscaling is refunded in full.
05Which preset should I pick?
Match it to the footage: aigc for AI-generated video, short_series for short dramas, ugc for footage shot by users, old_film for old film, and common for anything else. The preset is required for bytedance and doesn't change the price.
06Why won't my topaz result play in the browser?
topaz returns HEVC (h265), which some browsers can't play directly. Show preview_url, a compressed 600-pixel-wide h264 version, on web pages, and keep outputs[].url for the full-quality file.
07How long does upscaling take?
A 10-second clip upscaled to 1080p takes about two minutes with topaz and about two and a half with bytedance. Longer clips and higher resolutions take longer, and so does a busy service, so a webhook_url is the easiest way to know when it's done.
08Can I try it online before integrating?
Yes. Sign in with Google and use the studio at the top of this page: it runs the same model at the same prices as the AI Video Upscaler API, shows the estimated cost before you start and draws on the same balance.
Start Using the AI Video Upscaler API
Upscale a clip in the studio above, then bring the same AI video upscaler into your product with a single API call.