How the tiers work

• Lite — no download

  Classical, in-browser tools that run instantly on any device (even phones): filters, adjustments, smart-crop, quality score, A/B variations, classical denoise / deblur / colorize / face cleanup, and the procedural text-to-image composer. Nothing to download.

• Standard — ~400 MB

  Real ML models that download once then run offline: background removal, CLIP tags / categorize / similarity, captioning, OCR, document Q&A, depth, Whisper subtitles, and pose / segmentation. Runs on most laptops; a GPU helps, with WASM fallback.

• Pro — ~2 GB+ (WebGPU)

  GB-scale generative + restoration on a WebGPU GPU: SD-Turbo text-to-image, GFPGAN face restoration, plus bring-your-own ONNX model URLs. Falls back to the classical path where unsupported.

Models load from the Hugging Face CDN to your browser cache on opt-in — only model weights, never your files. Manage tiers, see what your device can run, and delete downloaded models anytime from the Model Manager (the database icon, top-right). See the AI Suite overview for details.