Flux Kontext Dev Ultra Fast

Flux Kontext Dev Ultra Fast

Playground

FLUX.1 Kontext Dev Ultra-Fast is an open-source image-to-image model that edits images from text prompts with open weights and code. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

FLUX Kontext Dev Ultra Fast — wavespeed-ai/flux-kontext-dev-ultra-fast

FLUX.1 Kontext Dev Ultra Fast is a low-latency image-to-image editing model optimized for rapid iteration. Provide a source image and a natural-language edit instruction, and it performs targeted or global edits while aiming to preserve the original context when requested—ideal for interactive workflows, batch revisions, and quick creative exploration.

Key capabilities

• Ultra-fast instruction-based image editing from a single input image
• Strong preservation when you explicitly specify what must remain unchanged
• Works well for iterative edits: refine the same image across multiple passes
• Great for practical edits: color changes, background swaps, text edits, cleanup, and light style transforms

Typical use cases

• Fast retouching and cleanup (lighting/exposure, minor imperfections)
• Color/material edits (e.g., product variants)
• Background replacement for marketing creatives
• Text replacement on posters, packaging, UI mockups
• Rapid style experimentation with minimal turnaround time

Pricing

$0.02 per image.

Cost per run = num_images × $0.02 Example: num_images = 4 → $0.08

Inputs and outputs

Input:

• One source image (upload or public URL)
• One edit instruction (prompt)

Output:

• One or more edited images (controlled by num_images)

Parameters

• prompt: Edit instruction describing what to change and what to keep
• image: Source image
• width / height: Output resolution
• num_inference_steps: More steps can improve fidelity but increases latency
• guidance_scale: Higher values follow the prompt more strongly; too high may over-edit
• num_images: Number of variations generated per run
• seed: Fixed value for reproducibility; -1 for random
• output_format: jpeg or png
• enable_base64_output: Return BASE64 instead of a URL (API only)
• enable_sync_mode: Wait for generation and return results directly (API only)

Prompting guide

Use a clear “preserve + edit + constraints” structure:

Template: Keep [what must stay]. Change [what to edit]. Ensure [constraints]. Match [lighting/shadows/style consistency].

Example prompts

• Keep the subject’s face and pose unchanged. Replace the background with a clean studio backdrop. Match the lighting and shadow direction.
• Change the shirt color to navy blue. Keep fabric texture and wrinkles consistent.
• Replace the label text with “WaveSpeedAI”, keeping the same font style, size, and perspective. Do not modify anything else.
• Remove the objects on the table. Keep the table surface intact and realistic.
• Apply a soft cinematic grade with slightly warmer tones, without changing composition or identity.

Best practices

• Do one change per run for maximum control, then iterate.
• If the edit drifts, lower guidance_scale and strengthen the preserve clause.
• Fix seed for reproducible comparisons and stable iteration.
• Match output width/height to the input aspect ratio to avoid distortion.

# Submit the task
curl --location --request POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/flux-kontext-dev-ultra-fast" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}" \
--data-raw '{
    "num_inference_steps": 28,
    "guidance_scale": 2.5,
    "num_images": 1,
    "seed": -1,
    "output_format": "jpeg",
    "enable_base64_output": false,
    "enable_sync_mode": false
}'

# Get the result
curl --location --request GET "https://api.wavespeed.ai/api/v3/predictions/${requestId}/result" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}"