Image Editing

Edit images with the power of generative AI. WaveSpeed transforms the editing workflow by allowing you to modify existing images using simple text prompts. Remove unwanted objects, expand canvas borders, change lighting, or completely restyle a photo without manual masking.

Start Editing Images View API Docs

Generative Editing Suite

A complete toolkit for AI-driven image manipulation.

1. Generative Fill (Inpainting)

Powered by FLUX Fill. Select any region of your image and describe what should replace it. The model generates photorealistic content that seamlessly blends with the surrounding pixels, matching lighting, perspective, and texture. Best for product photography retouching, creative compositing, and removing/replacing objects. Try the Image Eraser for quick object removal, or use FLUX Kontext for advanced prompt-guided editing.

2. Generative Expand (Outpainting)

Powered by Outpainting models. Extend your image beyond its original borders in any direction. The AI imagines and generates new content that naturally continues the scene — expanding landscapes, adding headroom to portraits, or converting vertical shots to widescreen. Best for social media repurposing (portrait to landscape), print design, and creating panoramic compositions from single shots.

3. Instruction Editing

Powered by Pix2Pix / InstructPix2Pix. Edit images using natural language instructions like "make it sunset," "add snow to the ground," or "change the dress to red." No masking or selection required — the model understands spatial context and applies changes only where appropriate. Best for rapid iteration, A/B testing visual concepts, and batch processing marketing assets. See our Best Open Source Image Models roundup for more options.

4. Smart Backgrounds

Combine Background Remover with generative fill to instantly swap backgrounds. Remove the original background with one click, then describe or upload the new environment. Perfect for e-commerce product shots (studio white to lifestyle scenes), real estate virtual staging, and social media content creation at scale. Available on WaveSpeed.

Transformation Examples

See how AI editing solves common visual challenges.

Product Photography

Remove distracting backgrounds, swap studio setups, and retouch products in seconds. Transform a single photo shoot into dozens of variations for A/B testing on e-commerce listings.

Real Estate Virtual Staging

Fill empty rooms with virtual furniture and decor using generative inpainting. Change wall colors, add rugs, or swap lighting fixtures without any physical staging costs.

Social Media Repurposing

Use outpainting to extend vertical portraits into landscape banners, or instruction editing to change seasons, time of day, or weather across an entire batch of marketing assets.

Creative Compositing

Combine multiple editing tools in one workflow: remove an object, fill the gap with AI-generated content, then upscale the final result for print-ready quality using the Image Upscaler.

Q & A

How is AI Image Editing different from Photoshop?

Traditional editing requires manual selection, cloning, and color adjustment. AI Image Editing uses Generative Models to create new pixels based on text. You don't "draw" the change; you "describe" it, and the AI renders it for you.

Does editing reduce image quality?

No. Our advanced inpainting models work at high resolutions. For significant expansions or edits, we also recommend running a final Upscale pass to ensure all generated details are crisp and sharp.

Can I edit specific parts without a mask?

Yes. Using models like InstructPix2Pix, you can simply type "Change the red car to blue," and the AI will automatically identify the car and change its color without affecting the rest of the image.

Is the content unique?

Yes. Every generation is unique. Even if you use the same prompt twice, the AI will generate slightly different variations, giving you multiple creative options to choose from.

What file formats are supported?

You can upload JPEG, PNG, or WEBP images. The edited result is typically provided as a high-quality PNG to preserve transparency (if applicable) and detail.