If you've clipped a full session and ended up with six or seven raw recordings instead of one, you've probably felt the annoyance of converting them one at a time — load a file, wait, download, repeat. TrackPeel now handles this as a queue: drop in as many WebM/MKV files as you want, hit Convert all, and it works through them in order, with per-file progress and download buttons as each one finishes.
Why this wasn't just "loop the existing code"
The obvious approach — just run the same conversion function in a for loop — actually breaks in a subtle way once real files are involved. The conversion engine (FFmpeg compiled to WebAssembly) manages a chunk of virtual memory internally, and previous file data needs to be fully cleared before the next file starts, or you start accumulating memory usage across the queue until the browser tab runs out of headroom. That's exactly what happened in early testing: the first two or three files in a queue converted fine, and then things quietly slowed down or failed.
The fix was to make sure each file gets loaded, converted, and had its filesystem entries removed from the engine's virtual memory before the next one starts — genuinely sequential, not just visually sequential. It's less glamorous than the demo GIF makes it look, but it's the difference between "works for me with three files" and actually holding up for someone converting a folder of twelve.
What you can do with it now
- Drop in a whole session's worth of recordings and walk away instead of babysitting one conversion at a time
- Remove individual files from the queue before converting if you change your mind about one
- Keep working on other things in the same tab while a large queue works through the background — the conversion runs in a separate worker thread, so it doesn't lock up the page
What's still on the list
Right-clicking a folder and having every WebM/MKV inside it picked up automatically (instead of selecting files one at a time) is next up. If there's something specific slowing down your workflow that this update doesn't touch, the About page has a real inbox we read — feature requests genuinely shape what gets built next.