Flux Kontext Pro
FLUX.1 Kontext [pro] is Black Forest Labs' model for text-to-image generation and editing a reference image with text instructions. This record describes the black-forest-labs/flux-kontext-pro offer on Replicate; its input fields and prices belong to that hosted interface.
- Review Kontext Max
BFL positions Max for stronger prompt adherence and typography; compare its separately scoped hosted record using the same image and instruction.
- Compare other image-editing models
Browse image-editing models and review the inputs and examples for your editing task.
🚀Function Overview
Use Kontext Pro for targeted changes to an existing image, such as replacing a sign or changing clothing while preserving the composition. BFL now documents Kontext as a previous-generation family and recommends FLUX.2 for new projects. Pro and Max are hosted variants; the separately downloadable Kontext Dev checkpoint has its own license. The Sources section separates BFL's model guidance from Replicate's API contract, checked October 28, 2026. Listed run totals are historical snapshots, not a quality measure.
Key Features
- Transforms existing images using text instructions
- Designed to preserve a reference character across edits; inspect identity and details after each result
- Allows text replacement in images
- Performs style transfers and background swaps
- Replicate provides safety_tolerance controls; its schema limits this setting to 2 when an input image is supplied
- For a sign edit, try: Replace 'OPEN' with 'CLOSED', keeping the lettering style and background unchanged. This is an unexecuted instruction example, not a measured result
- BFL reports that repeated editing can accumulate artifacts and that instructions may be missed; keep the original image and inspect each edit before continuing
Use Cases
- •Converting photos to artistic styles (e.g., watercolor, oil painting)
- •Modifying hairstyles, clothing, or accessories in images
- •Changing text on signs or posters
- •Swapping backgrounds while preserving subjects
- •Creating character-consistent variations from a single image
⚙️Input Parameters
prompt
stringText description of what you want to generate, or the instruction on how to edit the given image.
input_image
stringImage to use as reference. Must be jpeg, png, gif, or webp.
aspect_ratio
stringAspect ratio of the generated image. Use 'match_input_image' to match the aspect ratio of the input image.
seed
integerRandom seed. Set for reproducible generation
output_format
stringOutput format for the generated image
safety_tolerance
integerSafety tolerance, 0 is most strict and 6 is most permissive. 2 is currently the maximum allowed when input images are used.
💡Usage Examples
Example 1
Input Parameters
{
"prompt": "Make this a 90s cartoon",
"input_image": "https://replicate.delivery/pbxt/N55l5TWGh8mSlNzW8usReoaNhGbFwvLeZR3TX1NL4pd2Wtfv/replicate-prediction-f2d25rg6gnrma0cq257vdw2n4c.png",
"aspect_ratio": "match_input_image",
"output_format": "jpg",
"safety_tolerance": 2
}Quick Actions
Technical Specifications
- Hardware Type
- Not specified
- Run Count
- 1.9M
- Commercial Use
- Supported
- Pricing
- $0.04 per output image
- Platform
- Replicate
Sources
- Black Forest Labs: Kontext variants, current positioning and Dev license
- Black Forest Labs: reference-image and text-editing instructions
- Black Forest Labs: May 29, 2025 launch and documented failure cases
- Replicate: Kontext Pro hosted input and output schema
- Replicate: Kontext Pro offer and commercial-output terms