TL;DR
We've expanded Fiddl.art's model lineup with Imagen 4 Ultra, GPT-Image-1, and Recraft while upgrading our upscaling to Topaz for better detail preservation. Behind the scenes, we've streamlined our creation system and renamed Face Forge to Forge to reflect its broader custom model capabilities. Error messages are now more helpful, and we've laid groundwork for future features like inpainting and video generation.
More Models, More Possibilities
Our latest update brings several new AI models to Fiddl.art, each offering unique strengths for different creative needs:
- GPT-Image-1: Excels at interpreting complex text prompts with precision
- Recraft: Delivers design-focused outputs ideal for layout-aware content
- Photon: Produces clean, high-resolution images with natural tonal quality
- Imagen 4 & Imagen 4 Ultra: Provide professional-grade realism and exceptional detail
Each model now features clear pricing and supported aspect ratios, making it easier to choose the right tool for your project. Explore these options in our Models catalogue to find your perfect creative match.
Upscaling That Respects Your Image
We've significantly improved our upscaling technology by switching from Clarity Upscaler to a Topaz-based solution. The previous upscaler sometimes altered facial features and fine details too aggressively. Our new implementation maintains structural integrity while enhancing resolution, particularly important for portraits and detailed artwork.
The comparison demonstrates how Topaz preserves facial features and texture details that other upscalers might compromise. This upgrade ensures your AI-generated art maintains its original character while gaining enhanced clarity.
Behind the Scenes: A Cleaned-Up Core
Supporting these new models required substantial backend improvements. We refactored our image generation system to create a more consistent and extensible architecture:
- Standardized input validation and handling procedures
- Reduced model-specific tweaks through a unified framework
- Separated billing, generation, and storage logic for better maintainability
These changes might not be immediately visible to users, but they create a stronger foundation for future features like AI video generation and advanced editing tools. The updates also benefit developers using our API documentation for integrated workflows.
Face Forge → Forge
We've renamed Face Forge to simply Forge to better reflect its expanded capabilities beyond facial recognition. This change aligns with our vision for custom model training that supports diverse creative applications including styles, objects, and brand identities.
The updated Forge interface provides more intuitive controls for training custom AI models, making it accessible for both beginners and advanced creators. This evolution supports our growing ecosystem where creators can share and discover custom models through our improved discovery surfaces.
More Useful Error Messages
We've enhanced our error handling system to provide clearer, more actionable feedback when things go wrong. Instead of generic failure messages, you'll now receive specific information about what caused the issue—whether you're using the web interface or working with our API.
This improvement benefits everyone from casual creators to developers building integrated solutions. Better error messaging reduces frustration and helps users understand technical limitations or requirements for specific operations.
What's Next
Our engineering team is already working on several exciting developments:
- Forge upgrades: Advanced training controls using Flux Pro Trainer for more precise model customization
- Flux Kontext integration: Natural language editing capabilities for existing images
- Improved discovery: Enhanced browsing and onboarding experiences for new users
- Video expansion: Building on our current video capabilities with additional models and features
These initiatives continue our commitment to making Fiddl.art the most comprehensive AI art creation platform available.
Wrapping Up
This update strengthens Fiddl.art's foundation while expanding creative possibilities through new models and improved processing. The Topaz upscaler upgrade ensures higher quality outputs, while our backend improvements enable faster iteration on new features.
Whether you're exploring AI models for the first time or you're an experienced creator developing custom models, these updates provide more tools and better reliability for your creative workflow.
Frequently Asked Questions
What makes Imagen 4 Ultra different from other AI models?
Imagen 4 Ultra delivers exceptional realism and detail resolution, making it ideal for professional applications where photographic quality matters. It particularly excels at rendering complex textures, lighting scenarios, and fine details that other models might simplify.
How does Topaz upscaling improve my AI art?
Topaz upscaling preserves the original character of your AI-generated images while enhancing resolution. It maintains facial features, texture details, and artistic elements that other upscalers might alter, resulting in more faithful enlargements of your creations.
Can I use Forge for things other than faces?
Absolutely! While originally developed for facial recognition, Forge now supports training custom models for various applications including artistic styles, product representations, brand elements, and specific objects. The platform offers guided workflows for different use cases.
How do I choose the right AI model for my project?
Consider your priorities: GPT-Image-1 for prompt adherence, Recraft for design layouts, Photon for natural aesthetics, or Imagen 4 Ultra for maximum realism. Our model comparison guide provides detailed guidance on selecting the best option for your specific needs.
Are these updates available through the API?
Yes, all new models and the improved upscaling are available through our API documentation. The error message improvements also apply to API responses, providing better debugging information for developers.
