v1.0 live! Celebrating with 50% off -limited offer
-

Understand Pruna Image generator

Complete Pruna Image Edit platform guide: reference image requirements, aspect ratios, pricing, prompt tips and best practices for precise image editing.

Introduction

A practical guide to using the Pruna Image Edit model in our Media Generator: options, costs, and best practices.

Pruna Image Edit is a fast, precise image-editing model that requires a reference image. In our platform, it supports:

  • Image editing — Modify existing images via prompt
  • Multiple reference images — Up to 10 references
  • Aspect ratio control — Or match the input image
  • Low cost — One of the lowest-cost image models

Use it when you need to edit or transform an existing image (change background, style, elements) with high fidelity.

Where to Use Pruna Image Edit

| Location | How | |----------|-----| | Media Generator | Tools → Media Generator → Create Job → choose "Pruna Image Edit" → upload at least one reference image | | Image projects | Create a project, add reference images, and select Pruna Image Edit |

Requirements

  • Reference image required — At least one input image is mandatory
  • Reference defines the base content; the prompt describes the changes

Platform Options

Aspect ratio

| Option | Use case | |--------|----------| | matchinputimage | Keep input aspect ratio (default) | | 1:1 | Square output | | 16:9 | Landscape | | 9:16 | Portrait | | 4:3 | Standard | | 3:4 | Photo |

Default:

  • match

input

  • image

Output format

  • PNG — Lossless
  • JPG — Smaller file size

Default:

  • PNG

Reference images

  • Required — Minimum 1 image
  • Up to 10 reference images supported
  • Use multiple refs for complex edits or style blending

Costs (credits)

Pruna Image Edit uses flat per-image pricing:

| Item | USD | Credits | |------|-----|----------| | Per image | $0.01 | 0.03 |

\Credits computed with default billing config. Check the model selector in the UI for current rates. One of the lowest costs among our image models.

Prompt guidelines

Do

  • Describe what to change — e.g., “change background to beach,” “add snow”
  • Be precise about edits — “replace sky with sunset,” “remove the person”
  • Use clear, action-oriented language

Avoid

  • Vague instructions (“make it better”)
  • Asking for completely unrelated new content

Good practices

  1. Use clear reference images — Sharp, well-lit inputs give better edits.
  2. Be specific in the prompt — Describe the exact change you want.
  3. Match or choose aspect ratio — Use matchinputimage to preserve proportions.
  4. Use multiple references — Up to 10 for complex edits.

Limits and constraints

| Constraint | Value | |------------|-------| | Reference images | Required, 1–10 per generation | | Max images per request | 5 |

Troubleshooting

| Issue | What to try | |-------|-------------| | Missing reference warning | Upload at least one reference image | | Unexpected edits | Simplify and clarify the prompt | | Model not available | Confirm Replicate API token is configured |