LibraTool

Hi Fazier! I built LibraTool because every "free" PDF or audio tool I tried wanted me to upload private files, make an account, or wait out a daily limit. So every tool here runs in the browser: your files stay on your device. That includes the heavier ones, like Whisper speech-to-text and background removal, which run as on-device AI models. It now has 317 tools in 81 languages. I'd love to hear which tool you'd want next, or anything that doesn't work the way you expect.

The interesting engineering here is that the heavy jobs stay on-device. Running Whisper and ffmpeg (via WASM) locally is a real tradeoff, not just a privacy bullet point. Two questions. How does video conversion behave once a file goes past the browser's memory ceiling - is it chunked, or does it hand you off to something else? And can a sceptical user verify the never-uploaded claim directly, for example by watching the network tab while a tool runs?

Keeping Whisper and background removal fully on-device is the right call for the use cases you list, like transcribing a private interview or stripping GPS data from photos. How does the in-browser Whisper handle longer recordings on a low-memory laptop? I'd be curious whether there's a size guideline before it gets slow.

The omnitool framing is earned — 317 tools all running client-side is a genuine engineering feat, not a feature list. ffmpeg compiled to WASM plus on-device Whisper means the "your files are never uploaded" claim is verifiable, which is the whole point for contracts, interviews, anything personal. Most "free file tools" are upload farms with ads; this is the opposite bet. Congrats on the launch.

The GPS-removal use case would benefit from a before-and-after metadata report: show which EXIF fields were removed and which were retained, especially orientation and color profile. That would help people confirm location data is gone without accidentally changing how their photos display. Is that report available?

The never-upload angle is the part that sold me — most free PDF tools quietly want an account or a cloud round-trip, and stripping GPS before posting or merging a legal PDF is exactly when you don't want that. Two practical questions: is there a rough file-size ceiling where WASM ffmpeg starts to crawl, and does on-device Whisper ship as a one-time model download or per session? A small 'processed locally, 0 bytes sent' indicator would make the claim easy to verify mid-task.

For the HEIC-to-JPG use case you mention, how do you handle Display P3 photos and EXIF orientation? A note about whether conversion preserves the color profile or converts to sRGB would help people avoid unexpected color shifts when preparing photos for the web.

Running Whisper speech-to-text right in the browser is a big deal for anyone transcribing interviews with sensitive content. The tool for stripping GPS data from photos before posting is a lovely privacy win too. Which of the 317 tools has turned out to be the most popular so far?

Running Whisper and ffmpeg locally in the browser is the right call for this kind of tool. Transcribing an interview without uploading the audio anywhere is exactly what I would want for client files. Curious how it handles longer files on mobile: is there a practical size limit for video conversion before the tab runs out of memory? A small note on the page about that would save people a failed attempt.

Thats cool, I'm gonna use this

That's cool, I'm gonna use it