Frequently asked questions
The short answers, and where in the site each one comes from.
Everything below is about the tools as they are built, not a policy. Where a number is involved it is the number in the code.
Why was my file skipped?
A file is not added to the queue for one of three reasons, and the toast says which:
- It is under 50 MB, on TS → MP4 only. Sizes that small are almost always a fragment or a failed recording rather than something worth converting. MOV → MP4 has no size floor, because a small QuickTime clip is a normal file.
- It is not a file the tool takes. TS → MP4 accepts
.ts,.mtsand.m2ts; MOV → MP4 accepts.movand.qt; Video → Audio takes the common video containers (and audio files, which it simply re-exports). A same-named.jpg/.pngbeside a recording is used as its thumbnail and a.starsidecar becomes a flag on the row — neither is queued as a video. - You removed it earlier. Removing a file from the list is remembered by name for that tool, so re-adding the same name does nothing. The note above the queue shows how many names are remembered, and its “Clear list” button forgets them.
Nothing is skipped silently: a skipped file produces the toast naming the count and the reason.
Why does the engine reload in the middle of a batch?
FFmpeg runs as WebAssembly inside the page, and a wasm instance's memory only ever grows. Deleting the input and output from its virtual filesystem frees the files but not the ceiling the instance has already reached — so a long batch of large recordings eventually hits a memory trap. To avoid that, the engine is deliberately thrown away and reloaded, every six files or every ~1 GB of source video, whichever comes first.
The same recycle happens immediately after you press cancel, and after any fatal wasm error, because a trapped instance cannot be trusted with the next file. The reload is quick: the core is bundled with the site (not fetched from a CDN) and the browser has already cached it, so you see the “loading” phase rather than a download.
Is anything uploaded?
No. There is no upload endpoint to send anything to, and no server holding your files. Each file is read by the browser, written into FFmpeg's in-page filesystem, converted there, and handed back to you as a download from a blob URL that exists only in this tab. Close the tab and the output is gone unless you saved it.
The one third party the page loads is the Google AdSense script; what it and its cookies do is set out on the privacy page. Nothing about your files ever reaches it — the conversion code has no network call in it at all.
Are big batches safe?
Yes, with one practical limit. Files convert strictly one at a time, in name order, so a hundred-file queue is a queue rather than a hundred parallel jobs. The periodic engine recycle above is what keeps a long run from degrading.
The limit is per file: the source and the result both sit in browser memory while a job runs, so a single file over roughly 2 GB can fail no matter how much free disk you have. Splitting a very large recording is the workaround. Also keep the tab open and awake — the queue lives in the page, so a reload or a crash loses it, which is why the browser warns you before you leave with files still in the list. Download finished files as you go if the batch matters.
Does it work offline, and can I install it?
Every tool is a page with no server round-trip in its conversion path, so once the page and its FFmpeg core have loaded it keeps working without a network — on a plane, or on a locked-down laptop where a conversion site would be blocked. The site also ships a web app manifest, so a browser that offers “Install” can put it in its own window like an app. Installing changes nothing about the privacy story: it is still this page, running on your machine.
What if the output will not play, or the conversion fails?
A failed row keeps the error and the FFmpeg log, and the file is left in the list so you can retry it with a different mode. “Auto” copies streams that are already compatible and re-encodes only what has to change; “Stream copy” is the fast, lossless option that fails when the codec inside cannot live in an MP4; “Re-encode” always works but is much slower. If a file fails in all three, it is usually damaged at the source.
If something looks wrong that this page does not explain, email jacoblbagent@gmail.com with the file's codecs (the row's log has them) — the contact page lists what to include.
Where can I read more?
Each tool has a guide below it covering its formats and modes, the about page explains why the site is built this way, and the terms and privacy pages cover the rest.