Recover 3

Generative diffusion model that excels at recovering rich, natural detail

Model Overview

Recover 3 is a generative image recovery model designed to reconstruct missing detail in low-resolution, damaged, or out-of-focus images, with a particular strength in restoring realistic facial features and natural textures. Now in its third generation, Recover 3 delivers significantly improved detail reconstruction, texture fidelity, and realism compared to earlier versions.

Recover intelligently rebuilds hair, skin, fabric, foliage, and other organic detail where original image information is limited or lost. It is especially effective for old photographs, heavily compressed images, and small or blurry faces, producing results that feel photographic rather than artificial.

Key Capabilities

• Reconstructs missing facial features and fine textures with improved realism and stability over earlier Recover models.

• Prioritizes believable skin, hair, and material textures while avoiding over-sharpened or plastic artifacts.

• Realistic and Creative recovery options to balance fidelity versus detail generation, depending on source quality.

Model Pricing

Output MP	1	4	8	16	24	32	40	50	64	100
Credits	1	1	2	4	6	8	10	13	16	25

Getting Started

Endpoint

• https://api.topazlabs.com/image/v1/enhance-gen/async

  • model — "Recover 3" or "Natural Enhance"

Endpoint-Specific Parameters

• face_enhancement — boolean

  • Leverages our face recovery model to enhance faces.

• face_enhancement_strength — decimal between 0 and 1 (required if face_enhancement=true)

  • Sets the strength of the face recovery.

• face_enhancement_creativity — decimal between 0 and 1 (required if face_enhancement=true)

  • Choose realistic or creative face recovery.

• subject_detection — "foreground", "background", or "all"

  • Choose where enhancements are applied.

• prompt — string, default ""; max 1024 char

  • A description of the resulting image you are looking for. The model responds more to a descriptive statement versus a directive one. For example, use the phrase "girl with red hair and blue eyes" instead of "change the girl's hair to red and make her eyes blue"

• autoprompt — boolean (if enabled, ignores value given to prompt)

  • Whether you want to use our state-of-the-art auto-prompting model to auto-generate a prompt for you.

• creativity — integer between 1 and 9, default 3

  • Lower creativity values maintain the highest fidelity to the original image. Higher values take more liberties and provide more creative results for specific details.

• texture — integer between 1 and 5 (recommend 1 at low creativity, 3 at higher creativity)

  • Add texture to the image. Recommend setting texture to 1 for at a low creativity level, and 3 for more creative results at a higher creativity level.

• sharpen — decimal between 0 and 1

  • Slightly sharpens the image.

• denoise — decimal between 0 and 1

  • Reduces noise in the image.

• detail — boolean, default false

  • Adjusts the level of added detail after rendering.

• detail_strength — decimal between 0 and 10 (required if detail=true)

  • Sets the strength of the detail.

These parameters should be appended as key-value fields in the form-data request bodies (i.e. alongside the model, output_height, and face_enhancement etc. fields). Any parameters not explicitly provided are automatically configured by our auto-parameter model. Extra parameters provided that are not supported are ignored.

Model-Specific Parameters

• enhancement_strength: float, 0.0 <= x <= 10.0, default = 5.0

Resolution Limits

• input — 24 MP

• output — 384 MP

Examples

Click to view

Click to view