Does this page upload or process my video?
No. It only writes command text in your browser. You run the command on your own computer, where ffmpeg reads and writes the files.
Every option explained.
Pick a job such as trimming a clip, compressing to fit an upload limit, extracting audio, or making a GIF, and get a ready-to-run ffmpeg command with every option explained. Paste a command you found elsewhere to have it decoded the same way.
Loading the toolbox.
You do not have to take our word that your data stays here. Developer tools show every request a page makes, so you can watch this page work and confirm that nothing is sent. The check takes under a minute.
F12, or Cmd+Option+I on a Mac, and select the Network tab.Only two things can add a line to that log, and neither carries your data. Submitting the feedback form sends the vote and comment you typed. Typing characters in a script the page has not displayed yet, such as Cyrillic or Greek, can make the browser fetch the matching font file, and that request names only the font.
The production site also sends a Content-Security-Policy header that tells your browser to refuse connections to anywhere except this site and the anonymous analytics endpoint, so a coding mistake could not quietly send your data elsewhere.
ffmpeg is the open-source program behind many video tools. It can convert, cut, compress, and filter almost any audio or video file, but its options are terse and order-sensitive, so most people copy commands from forum answers without knowing what each part does. This page turns that around: you describe the job, and it writes the command and explains it, option by option.
It covers the tasks people search for most often:
The builder never runs ffmpeg and never sees your files. It only writes text, in this tab, for you to run on your own machine.
Paste any ffmpeg command into the decoder and each option is listed with a plain-English explanation and the file it applies to. That last part matters more than it looks. ffmpeg reads its options in order: an option applies to the next -i input or the next output file name that follows it. The same -ss 30 means fast input seeking before -i and slow, exact output seeking after it.
The decoder tracks that position for you and warns when something is in the wrong place, for example a video filter written before the input, or options left after the last output file, which ffmpeg ignores. It also flags combinations that fail, such as -vf filters together with -c:v copy, an output file that is also an input, and the -y flag, which overwrites an existing output file without asking and is marked as destructive. Shell quoting is undone the way a shell does it, so a file name in quotes with spaces is read as one name. Options the decoder does not know are shown and marked rather than guessed at.
File size is bitrate multiplied by duration, so a size limit and a length give you a bitrate budget. The page uses these formulas, which are also shown in its How this is calculated fold:
total_kbps = size_bytes x 8 / duration_seconds / 1000video_kbps = floor(total_kbps x (1 - margin / 100) - audio_kbps)A worked example: a 25 MB limit is 25,000,000 bytes. For a 60-second clip that allows 3,333.333 kbps in total. Keeping a 2% margin leaves 3,266.667 kbps, and taking off 128 kbps for audio leaves 3,138 kbps for video once rounded down, so the commands use -b:v 3138k. If the limit is 25 MiB instead, which is 26,214,400 bytes, the same clip can use 3,297 kbps for video.
The margin covers the container's own data, such as headers, the index, and timestamps. Its size depends on the format and the length, so the 2% default is an assumption you can change, not a measured figure. If the result comes out over the limit, raise the margin. When the audio alone would use up the budget, or video would get less than 100 kbps, the page says so instead of writing a command; that floor is this tool's own cut-off, not a standard.
A quality target such as CRF makes every scene look about the same, but you cannot know the final size until the encode ends. To hit a size, you need a bitrate target instead, and a single bitrate pass spreads bits evenly, wasting them on easy scenes and starving hard ones. Two-pass encoding fixes that:
/dev/null on macOS and Linux, NUL on Windows) and is thrown away.Run both commands from the same folder, in order, and delete the ffmpeg2pass log files when you are done. If you do not know the exact length of the video, the page also gives you the ffprobe command that prints it.
With -c copy, ffmpeg moves the compressed audio and video into the new file untouched. It is very fast and loses no quality, which makes it right for changing a container or cutting a rough clip. The catch is that a copied cut can only start on a keyframe, so a trim may begin a moment early or open on frozen frames, and no filter, resize, or quality change is possible.
Re-encoding decodes every frame and compresses it again. It is slower and loses a little quality each time, but cuts are exact and every filter is available. The builder picks the right mode for each task, explains the trade-off where you have a choice, and warns when a pasted command asks for both at once.
CRF, the Constant Rate Factor, is a quality target. According to the x264 and x265 documentation, the scale runs from 0 to 51, where lower means better quality and a bigger file; x264 defaults to 23 and x265 to 28. The page uses those defaults unless you change them.
The preset trades encoding time for compression. Slower presets, such as slow or veryslow, find a smaller file at the same quality; faster ones finish sooner with a bigger file. The default, medium, is a sensible middle. For H.265 output the builder adds -tag:v hvc1, which Apple players need to recognize the track, and for MP4 output it adds -movflags +faststart so web playback can begin before the whole file downloads.
The commands only read your input files and write new ones; none of them edits a file in place. Two things deserve care. First, ffmpeg asks before it replaces an existing output file, but a command with -y replaces it without asking, so the page marks -y as destructive. The only built command that uses it is pass 1 of a two-pass encode, where it lets ffmpeg write to the null device. Second, writing the output over the input destroys the original, so the decoder warns when the two names match. Test on a copy of an important file first.
Other command builders in this lane cover tar archives, git recovery, OpenSSL certificates, systemd services, and Dockerfiles.
Elsewhere in the toolbox: UUID, .gitignore, CSP, .htaccess, and more on the 2toolbox home page. Part of the 2suite.io privacy-first ecosystem.
No. It only writes command text in your browser. You run the command on your own computer, where ffmpeg reads and writes the files.
They use options that work in current ffmpeg releases, 7.x and 8.x, and most also work in older versions. Hardware encoders and some codecs depend on how your ffmpeg was built.
A stream-copy trim can only start on a keyframe. Choose the re-encode method for a cut that is exact to the frame.
The container adds data the bitrate does not count, and encoders overshoot a little. Raise the overhead margin by a percent or two and encode again.
It tells ffmpeg to overwrite an existing output file without asking. It is convenient in scripts, but it can destroy a file you meant to keep.