VRAM Fit Check

Shows whether each model and quant fits your GPU/RAM, estimated tokens/sec, and the exact ollama pull tag, right on the model page.

See if a model fits your GPU before you download it. People browsing Hugging Face or the Ollama library who want to know if a model will fit their GPU before downloading it. WHAT YOU CAN DO • GPU fit badges: Adds a colored badge next to each quant or file on a Hugging Face or Ollama library model page showing whether it fits your GPU's VRAM (green fits comfortably, yellow tight, red won't fit). • Estimated tokens/sec: Shows a rough tokens-per-second estimate for each quant based on your selected GPU class and the model's parameter count. • One-click ollama pull tag: Shows the exact `ollama pull <model>:<tag>` command for the quant you're viewing, with a button to copy it to the clipboard. • GPU and VRAM presets: Pick your GPU from a list of common consumer and prosumer GPUs, or enter a custom VRAM/RAM amount. The choice is saved on your device for future visits. • On/off toggle: A checkbox in the popup turns fit badges on or off, so you can quickly compare the page with and without them. PRIVACY BY DESIGN • Everything stays on your device. No account, no ads, no analytics. • The extension contains no code that sends data to a server. • Does not run in Incognito windows. VRAM Fit Check is an independent product by PinPixels. Questions and bugs: https://github.com/pinpixels/canirun/issues Privacy policy: https://pinpixels.github.io/canirun/privacy.html