Settings
Storage
Checking…
Default behavior on this device
Heavy tools
These tools download additional model files on first use. On mobile or metered connections, you can control which ones cache automatically.
Extract text from images using optical character recognition.
Detect every face in a photo and blur it — runs entirely on your device.
Upscale low-quality audio to clean 48 kHz. Great for cleaning up phone recordings, Zoom calls, or old podcasts.
1 speech tool (Whisper-tiny for transcription). Downloads ~250 MB on first use — runs entirely on your device.
2 image-to-text tools (caption / describe / extract). Downloads ~350 MB on first use — runs entirely on your device.
Tool group: gdal. Downloads ~40 MB on first use.
32 audio and video conversion tools powered by ffmpeg.wasm. Downloads ~30 MB on first use.
4 AI-powered image tools (background removal, upscaling, OCR Pro, image similarity). Downloads ~359 MB on first use — runs entirely on your device.
4 natural language processing tools. Downloads ~278 MB on first use — runs entirely on your device.
Translation model (400 MB). Only needed as a fallback — Chrome 131+ and Firefox use their built-in translator automatically.
"Enabled" means this tool's model will be cached when you use it. Actual cache contents and total disk usage are shown in the Storage panel above; clear them there.