TLDR
Fiddl.art now supports instant sign-in via Privy, replacing our previous email-code system with reliable magic links and social logins. We've also polished the gallery with a tighter grid and darker backdrop to make artwork pop, and migrated to a robust OpenAPI backend for clearer errors and easier third-party integration. These updates set the stage for future Web3 functionality and smoother multi-platform development.
Platform Improvements Walkthrough
Privy Sign-In
We retired our fragile, home-grown email-code login system due to frequent code delivery issues and limited OAuth options. Privy solves these problems with rock-solid email magic links, instant OAuth through major platforms, and a clear path to future Web3 wallet integrations. This reduces friction for users and minimizes authentication headaches for our engineering team.
Gallery Polish
The Browse page now features a tighter mosaic layout with square edges and a black background that makes colors more vibrant. These styling improvements reduce page weight and help every creation stand out, especially in low-light environments. The updated design enhances the overall experience when exploring Fiddl.art's growing collection of AI-generated artwork.
OpenAPI Backbone
Our original tRPC-based API served us well initially, but as Fiddl.art's codebase grew in complexity, we encountered type inference issues and unreliable builds. We migrated to OpenAPI to provide explicit, stable schemas and enable auto-generation of type-safe client libraries across multiple languages and platforms. This makes the Fiddl.art API easier for third parties to integrate and builds a stronger foundation for future development. Explore our OpenAPI specification to see the improved structure.
What's Next
Our public Trello board outlines the complete roadmap, but here are some key initiatives we're excited about:
Video Support
With advancements in video generation technology, we're preparing to add video support to Fiddl.art. We plan to make it fun and accessible by allowing users to animate existing creations or generate videos from text prompts. This expansion will complement our existing image generation capabilities and open new creative possibilities.
Additional Image Models
We're evaluating several promising image models for integration:
- Google Imagen 4 for enhanced realism and detail
- Flux Kontext for advanced editing capabilities (requiring additional UI work)
- Luma Photon for diverse stylistic options
These additions will give creators more tools and styles to work with when generating AI art.
Enhanced Custom Models Functionality
While our Face Forge feature already enables custom model training, we want to clarify that custom models can serve purposes beyond facial recognition. We're improving the UI to better communicate this versatility and developing enhanced sharing functionality. Soon you'll be able to browse models published by other creators and earn Fiddl points when others use your models. We're also adding support for training Flux Pro models alongside Flux Dev, which will improve output fidelity and accuracy.
Improved Browse Experience
The current Browse page shows a chronological list of platform creations with basic filtering options. As our collection grows, we're developing a more dynamic display system that will showcase popular images alongside randomized selections. This approach ensures you see fresh content and discover creations from before your account creation, providing a more diverse browsing experience with less scrolling.
Frequently Asked Questions
What is Privy authentication and how does it benefit Fiddl.art users?
Privy provides passwordless authentication through magic links and social logins, making sign-in faster and more reliable. It replaces our previous system that had delivery issues and offers a foundation for future Web3 functionality like wallet-based authentication.
How does the OpenAPI migration improve the Fiddl.art experience?
The OpenAPI backend provides clearer error messages, stable API schemas, and enables automatic generation of type-safe client libraries. This makes our API more reliable for internal development and easier for third-party developers to integrate with Fiddl.art's capabilities.
When will video generation be available on Fiddl.art?
We're actively working on video support and plan to release it in phases. The initial implementation will focus on animating existing creations and generating videos from text prompts, with more advanced features following based on user feedback and technological developments.
How can I create and share custom models on Fiddl.art?
Our Face Forge feature currently supports training custom models, primarily focused on facial recognition. We're expanding this to support more use cases and improving the sharing interface to let you browse others' models and earn points when your models are used.
What improvements are coming to the Browse page?
We're developing a smarter Browse algorithm that combines popularity signals with randomization to show both trending and diverse content. This will help users discover more artwork with less scrolling and ensure newer users see creations from before they joined the platform.