Upscale images with Topaz Wonder & Recover models. Reconstruct realistic detail, textures, and faces at high resolution. Try free — pay per generation.
Upscale an image with Topaz's Wonder/Recover generative models, reconstructing realistic fine detail, textures, and faces at high resolution.
About this model
Topaz Upscale Image Generative reconstructs realistic fine detail — skin texture, fabric, foliage, faces — using Topaz's Wonder and Recover model families. Unlike a purely creative upscaler, Wonder models aim to recover what a higher-resolution capture would plausibly have looked like, with dedicated face-enhancement controls and subject-detection to focus enhancement on foreground, background, or the whole frame. Nine model variants (Wonder 3.5/3/2, Recover 3, Standard MAX, Redefine, Recovery V2/Recovery) cover everything from general photo restoration to heavier reconstruction of badly degraded sources. Pairs well with Topaz Upscale Image Precision for a lighter-touch faithful pass, or Topaz Upscale Image Creative for a more stylized result.
Cost analysis
| Provider | Cost | Notes |
|---|---|---|
| muapiapp | $0.08+ per started 4-8 megapixels of output depending on model (plus margin), billed dynamically | Pay-per-use, no subscription. Cost is calculated from the real output resolution/duration at request time. |
| Fal.ai | Not available | Not offered as a standalone public API by muapiapp's upstream infrastructure partner. |
| Replicate | Not available | Not currently listed on Replicate. |
Pay-per-use, no subscription. Cost is calculated from the real output resolution/duration at request time.
Not offered as a standalone public API by muapiapp's upstream infrastructure partner.
Not currently listed on Replicate.
** Competitor pricing is estimated based on similar model architectures and usage tiers.
Configuration schema
| Parameter | Type | Description | Default |
|---|---|---|---|
| Image URL | string | Input image to upscale. | https://d3adwkbyhxyrtq.cloudfront.net/webassets/imagemodels/topaz-upscaler-in.jpg |
| Model | Enum (9 options) | Topaz generative upscale model variant. | Wonder 3 |
| Upscale Factor | number | Factor to upscale the image by. | 2 |
| Crop To Fill | boolean | Crop the output to exactly fill the target aspect ratio. | false |
| Output Format | Enum (2 options) | Output image file format. | jpeg |
| Subject Detection | Enum (3 options) | Which region of the image to prioritize enhancement on. | All |
| Face Enhancement | boolean | Enable dedicated face enhancement. | true |
| Face Enhancement Strength | number | Strength of face enhancement. | 0.8 |
| Face Enhancement Creativity | number | Creativity applied to face enhancement. | 0 |
| Enhancement Strength | Enum (3 options) | Overall enhancement strength tier. | - |
| Creativity | int | Creative detail strength (1-6). | - |
| Texture | int | Texture detail strength (1-5). | - |
| Prompt | string | Optional text prompt to guide generative detail. | - |
| Autoprompt | boolean | Automatically generate a guiding prompt. | - |
| Sharpen | number | Sharpening amount (0-1). | - |
| Denoise | number | Denoising amount (0-1). | - |
| Detail | number | Fine-detail amount (0-1). | - |
Input image to upscale.
https://d3adwkbyhxyrtq.cloudfront.net/webassets/imagemodels/topaz-upscaler-in.jpgTopaz generative upscale model variant.
Wonder 3Factor to upscale the image by.
2Crop the output to exactly fill the target aspect ratio.
falseOutput image file format.
jpegWhich region of the image to prioritize enhancement on.
AllEnable dedicated face enhancement.
trueStrength of face enhancement.
0.8Creativity applied to face enhancement.
0Overall enhancement strength tier.
-Creative detail strength (1-6).
-Texture detail strength (1-5).
-Optional text prompt to guide generative detail.
-Automatically generate a guiding prompt.
-Sharpening amount (0-1).
-Denoising amount (0-1).
-Fine-detail amount (0-1).
-Developer documentation
Provide an image: Set image_url to a publicly accessible image URL.
Choose a model: Wonder 3 is the default general-purpose choice; use Wonder 3.5 for the newest model, or Recover 3/Recovery V2 for heavily degraded sources.
Tune face and detail controls: face_enhancement, face_enhancement_strength, creativity, texture, sharpen, and denoise all shape the reconstruction.
Optionally guide it: Add a prompt (or set autoprompt: true) to steer what detail gets reconstructed.
Submit and poll: POST to /api/v1/topaz-upscale-image-generative, then poll /api/v1/predictions/{request_id} or use webhook_url.
curl -X POST https://api.muapi.ai/api/v1/topaz-upscale-image-generative \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"image_url": "https://example.com/portrait.jpg",
"model": "Wonder 3",
"upscale_factor": 2,
"face_enhancement": true
}'
Frequently asked
It reconstructs realistic fine detail (textures, faces, foliage) while upscaling an image up to 4x, using Topaz's Wonder and Recover model families.
Wonder 3 (default) is a solid general choice; Wonder 3.5 is the newest revision; Recover 3 and Recovery V2 target heavily degraded or low-quality sources.
Yes, `face_enhancement` defaults to true with `face_enhancement_strength` at 0.8 — both are adjustable, and `face_enhancement_creativity` controls how much new facial detail is invented.
Yes — `subject_detection` accepts `All`, `Foreground`, or `Background` to focus enhancement on a specific region.
Wonder 3/3.5 bill per started 8-megapixel block of output; all other models bill per started 4-megapixel block, both based on the real output size after `upscale_factor`.
Yes, the optional `prompt` field used to steer reconstruction accepts up to 1024 characters.