Changelog
Improved

FFmpeg jobs run your command exactly as written

Send the FFmpeg command you would run locally. Rendobar runs it as written on FFmpeg 9.0 and returns every file it produces, with your own output names and ffmpeg's real errors.

API

Send the FFmpeg command you would run on your own machine. Rendobar fetches your inputs, runs your argv exactly as written on FFmpeg 9.0, and hands back every file the command produces. It behaves the way it does locally.

A terminal running an unchanged ffmpeg command, with FFmpeg 9.0 and a sha256 on the returned file
  • Any input. Audio, subtitles, concat lists, fonts, images or data files. Whatever your command reads, it gets, under the name you referenced.
  • Inputs FFmpeg generates itself. -f lavfi -i anullsrc, image sequences like frame_%03d.png, -pattern_type glob and the concat: protocol all work.
  • Every file you write, named as you wrote it. A single output keeps your filename. A command that only measures, like -af volumedetect -f null -, returns ffmpeg’s log in output.data.log. Output sent to - comes back as a file. Two-pass log files come back.
  • Multi-file outputs, no per-file size limit. HLS ladders, resolution sets and image sequences return complete, and every file in output.files carries its sha256.
  • ffmpeg’s own error messages. A command that fails reports ffmpeg’s first error line, such as Unknown encoder 'libfoo', so you can act on it.
  • The media type you produced. An mp3 reports mediaType: "audio", an image "image", a hash or log null.
  • FFmpeg 9.0 on every job, with the exact build in each job’s ffmpegVersion.

A command runs in its own folder and reads and writes files by the names you give them, so reference each input by name. Shell syntax such as | head or 2>&1 is refused at submit, with a message naming it, since no shell runs here.

The ffmpeg job docs cover inputs, outputs and errors.

Share