Flora Fauna

Flora Fauna AI, branded FLORA, operates a browser-based generative AI workspace built around an infinite, node-based canvas. Rather than training a single proprietary model, the platform aggregates a broad set of third-party text, image and video models so that users can chain them together in one workspace: a text node can refine a brief or analyze an image, feed an image node, which can in turn feed a video node. Sources describe text model options including Claude 3 Sonnet, GPT o4 mini/GPT-4o and Google Gemini 2.0 Flash; image models including Flux Dev and Flux Pro 1.1, Ideogram 2.0, Luma Photon, Recraft V3 and Stable Diffusion 3.5, plus Flux Canny, Flux Depth and Flux Redux conditioning and background removal; and video models including Hailuo Minimax, Kling 1.5/Pro 1.6/2.0, Luma Ray 2 and Ray 2 Flash, Pika, Runway Gen-3 Alpha, Tencent Hunyuan, Lightricks LTXV, Google Veo 2 and WAN 2.1. The company's own site additionally references Nano/Nano Pro and Veo 3.1 outputs.

The product is positioned around three stages: ideation across many models on one canvas, real-time team iteration, and scaling a single concept into large volumes of production-grade, on-brand assets. Collaboration features include shared projects, group work and a commenting system for feedback. Since late March 2025 users have been able to train custom styles by uploading roughly 20-30 reference images, after which the style becomes selectable for generation. The site lists a canvas product, a vertical "Fashion Studio" offering (sketch-to-render, flatlay, ghostform, garment recolor, fabric swap, model maker, photo shoot, model try-on, 360 model video), a technique library, MCP/API access, docs, a trust center and status page.

Reviewers characterize FLORA as sitting between consumer chat-style generators and developer tooling such as ComfyUI, with a meaningful learning curve. A hands-on review reported roughly 500 free credits for new users, generation times of at least about 30 seconds even on faster models, strong results on still-image and multi-step ad-campaign workflows, and weaker, comparatively expensive video output.