Launching the most efficient, fastest video upscaling on Mac — FlashVSR+'s one-step diffusion pipeline running at full GPU speed through vpipe's Metal backend.
🚀 How to use
Model — choose Turbo or Base and hit Fetch model (first time only).
You'll see per-file download status and a live progress bar.
Input Video — drop an mp4 / mov / mkv / webm file.
Settings — pick the target resolution (with aspect-lock / match-source-aspect),
output FPS, color fix, decoder (Tiny fast / Full VAE), temporal compression,
seed (or random), and whether to keep the source audio.
⚡ Upscale Video — the ⚡ logo flickers to life and the job runs in the
background; watch the phase, clip counter and log tail (the red Cancel stops it).
When it's done, the result appears in the Result tab.
Compare — drag the divider (or the slider) to sweep between original and
upscaled; Play keeps both in sync; use the seek bar to jump around.
Scroll to zoom, drag to pan, double-click to reset.
Library — all past outputs with previews and badges. Each card: ⬇ download,
⇄ compare, 🗑 delete (just that generation), or click to open it on the
Result page. ⟳ refreshes; 🗑 Clear wipes the whole library.
Posts · 8
New on Pinokio
Automatically discovered via GitHub
Update v3 :
Changelogs :
1. input image size for the uploaded image shown in the UI.
2. zoom,pan for the result tab .
3. model download progress bar in UI, background downloads support.
4. a paste from clipboard button( along with keyboard shortcut) to add images
5. option for random seed by default.
New Update with Multiple/batch import, Library tab!
Updated with Multiple/batch import support, New Library tab for viewing multiple images in a session.. Fastest and most Memory Efficient Seedvr2 implementation for Apple Silicon Mac..
New on Pinokio
Automatically discovered via GitHub
Update v4 is here! Lora Support, Hum to Song, bug fixes and more...
Changelogs : Lora support is here! Dedicated Hum to Song tab with the new lora. UI cleanups Code cleanups Fix env to use uv READ THE INFO TAB, for links to the supported loras.
UPDATE v3 : Cover Song, Better memory managent, Faster generations.
Cover Song support is here! Better Peak Memory optimisations (~6-8gb peak with 8 bit model ). Fix for model management. IMPORTANT NOTE : EXISTING USERS SHOULD DOWNLOAD the update.json from the repo and replace it locally and then run the update, as I've updated the python version(inorder to add support for the new cover song feature). Lora support coming soon...
You’re all caught up.