What “processed on your device” means
Where your file goes, what gets downloaded, and why large jobs can fail.
The code travels. Your file stays.
Your browser downloads this website and a processing engine. After you choose a file, that engine reads the file into browser memory and writes the result there. The tool does not upload the selected file to a conversion service.
The first run has a download
The media engine is much larger than the page itself and loads only when processing begins. The browser may cache it for later visits. Private browsing or clearing site data can remove that cache. Loading the page alone does not make the tool fully offline-ready.
Memory is the practical limit
The original data, decoded media, output, and processing engine can coexist in memory. A small compressed file can expand substantially. We limit input to 100 MB and total duration to 20 minutes, with an additional 150 MB estimated WAV-output limit. These boundaries reduce failures but cannot guarantee success on every phone or browser.
What we can measure
Optional usage analytics can report tool names, broad file-size ranges, duration and error categories. They do not include your file name or media contents. Download initiation is measurable; your browser generally cannot tell a website whether you actually saved the file, so confirmation is separate.
Put it to use
Compress an MP3 or trim your audio.
Technical reference: FFmpeg WebAssembly FAQ.