Introducing WaveSpeedAI Firered Image V1.1 Edit on WaveSpeedAI

Introducing FireRed Image Edit V1.1: Enhanced Identity Consistency and Multi-Element Fusion

WaveSpeedAI is thrilled to announce the availability of FireRed Image Edit V1.1, the latest upgrade to Xiaohongshu’s acclaimed image editing foundation model. Building on the already impressive V1.0 release, this new version delivers significantly improved identity consistency, more powerful multi-image conditioning, and domain-specialized editing capabilities that bring AI-powered image editing closer than ever to real-world creative production needs.

What is FireRed Image Edit V1.1?

FireRed Image Edit V1.1 is a diffusion transformer-based image editing model that lets you transform images using natural language instructions. Upload up to three reference images, describe the output you want in plain English or Chinese, and the model generates a new image that intelligently combines elements, styles, and characters from your references.

The V1.1 release is not a minor patch—it represents a meaningful leap in the areas that matter most for professional creative work. The FireRed team focused their optimization efforts on four key dimensions: portrait consistency, multi-element fusion, stylized text rendering, and portrait makeup effects. The result is a model that maintains state-of-the-art open-source performance on major benchmarks (Imgedit, Gedit, and RedEdit) while excelling in the specialized editing scenarios that professional creators encounter daily.

What’s New in V1.1

State-of-the-Art Identity Consistency

The most significant improvement in V1.1 is character identity preservation. When editing portraits—changing backgrounds, applying new styles, or compositing subjects into different scenes—the model now maintains facial features, expressions, and distinguishing characteristics with markedly higher fidelity. This is powered by a differentiable consistency loss mechanism that extracts regions of interest from both predicted and ground truth images, minimizing identity discrepancy throughout the generation process.

For creators working on character-driven content, this means you can restyle, repose, and recompose subjects without the identity drift that plagued earlier models.

Multi-Element Fusion at Scale

V1.1 takes multi-image composition to another level. While the base model supports up to three direct reference images, the built-in Agent module enables compositions with 10 or more elements through automatic ROI detection, cropping, and stitching. Complex compositions that previously required careful prompt engineering and multiple attempts now work reliably with straightforward descriptions.

Comprehensive Portrait Makeup

A new specialty in V1.1 is its portrait makeup system, offering dozens of styles ranging from professional beauty retouching and skin tone enhancement to creative and theatrical looks. This capability is particularly valuable for beauty brands, content creators, and editorial teams who need consistent, high-quality makeup applications across large sets of images.

Stylized Text Reference

V1.1 introduces improved text style rendering with dedicated LoRA support, enabling high-fidelity typography preservation and stylized text effects. Whether you’re incorporating branded text elements or preserving existing text in edited images, the model handles it with greater precision than its predecessor.

Key Features

• Natural language editing: Describe edits in plain English or Chinese with high prompt-following accuracy
• Multi-image references: Upload up to 3 reference images and combine elements by referencing them as “Figure 1”, “Figure 2”, etc.
• Style-aware generation: Excels at editorial looks, film grain, retro tones, and cinematic aesthetics
• Flexible output sizing: Choose preset aspect ratios (1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3) or custom dimensions from 256 to 1,536 pixels
• Built-in prompt enhancer: Automatically refines your descriptions for better results
• Character consistency: Preserve subject identity while transforming style, setting, or mood
• Photo restoration: Repair scratches, enhance faded details, and upscale old photographs without artificial artifacts

Real-World Use Cases

Editorial and Fashion Photography

Create stylized portraits with film grain, warm retro tones, and classic editorial aesthetics. Transform subjects across different settings, eras, or moods while keeping their identity intact—ideal for lookbooks, campaign imagery, and social media content at scale.

Virtual Try-On and E-Commerce

Combine a product image with a model photo to generate realistic composites showing how garments, accessories, or cosmetics look on different subjects. V1.1’s improved identity consistency ensures the model’s appearance stays natural and recognizable.

Creative Compositing

Merge elements from multiple reference images into cohesive scenes. Combine a character from one photo, a background from another, and a style reference from a third—the Agent-powered pipeline handles the visual harmonization automatically.

Beauty and Cosmetics

Apply professional makeup styles consistently across product photography or influencer content. The portrait makeup system handles everything from subtle retouching to dramatic creative looks with consistent quality.

Brand and Marketing Content

Rapidly iterate on campaign visuals by combining brand assets, products, and models into polished compositions. Generate multiple variations without returning to the studio, saving both time and budget.

Getting Started on WaveSpeedAI

Start using FireRed Image Edit V1.1 in minutes:

1. Visit the model page: Go to wavespeed.ai/models/wavespeed-ai/firered-image-v1.1/edit
2. Upload reference images: Add up to 3 images for style, character, or scene guidance
3. Write your prompt: Describe the desired output, referencing images by Figure number for precise control
4. Set size (optional): Choose a preset aspect ratio, enter custom dimensions, or leave empty to preserve the original proportions
5. Generate: Submit and download your edited image

Or integrate directly via the WaveSpeedAI API:

import wavespeed

output = wavespeed.run(
    "wavespeed-ai/firered-image-v1.1/edit",
    {
        "prompt": "Put the person from Figure 1 in a vintage film portrait style with warm retro tones and soft film grain",
        "images": [
            "https://example.com/portrait.jpg",
        ],
    },
)

print(output["outputs"][0])

At just $0.08 per image, FireRed Image Edit V1.1 delivers professional-grade editing at a fraction of the cost of manual work or traditional photo production.

Why WaveSpeedAI?

WaveSpeedAI provides the fastest, most reliable way to run FireRed Image Edit V1.1 in production:

• No cold starts: Requests begin processing immediately—no warm-up delays
• Fast inference: Optimized infrastructure delivers results in seconds
• Affordable pricing: Transparent pay-per-use pricing at $0.08 per image
• Production-ready API: Seamless integration into your applications and workflows

Start Editing with V1.1 Today

FireRed Image Edit V1.1 raises the bar for open-source image editing—delivering stronger identity preservation, more powerful multi-element fusion, and specialized capabilities for portrait, makeup, and text editing that meet the demands of professional creative workflows. Whether you’re building e-commerce experiences, producing editorial content, or automating marketing visuals, V1.1 gives you the precision and quality to do it at scale.

Try FireRed Image Edit V1.1 now on WaveSpeedAI at wavespeed.ai/models/wavespeed-ai/firered-image-v1.1/edit—no setup required, no cold starts, just instant access to state-of-the-art image editing.