Updated Sep 21, 2026

Utilities

Convert a media file or folder into another format

Convert one file or a folder of media and write the results next to each other.

Agent

overview

What ffmpeg does

You have a WAV, MP4, or a folder of those files, and the next node needs a different container or codec. Set -i for one file or --input-folder for a directory. The wrapper calls ffmpeg, writes into the output folder, and names each result after the source stem plus --output-ext. Default extension is mp4. --output-ext mp3 is the audio extract used in the smoke run.

Optional controls stay off the default canvas. --video-codec and --audio-codec map to -c:v and -c:a. --extra passes a raw ffmpeg argument string. --cuts-file reads start, end, and output name per line. --subtitle-file or --subtitles-folder can burn captions when you already have an ASS, SRT, or VTT next to the video.

The node converts files you already have. It does not download a stream or invent a codec the image lacks. A completed run wrote tone.mp3 (1413 bytes) from a one-second fixture WAV. Treat an empty folder as a failed convert, not a silent success.

source github.com/FFmpeg/FFmpeg

use cases

Where ffmpeg fits

Turn a WAV into an MP3

Set -i and --output-ext mp3. The folder holds stem.mp3 for the next node.

Convert a folder of clips

Pass --input-folder. The wrapper walks common audio, video, and image extensions and writes one output per file.

Cut segments from a cuts file

Set --cuts-file with start, end, and output name per line. Pair it with -i or --input-folder.

Burn a subtitle file into a video

Set --subtitle-file when you already have ASS, SRT, or VTT. Do not expect the node to fetch captions.

reference

ffmpeg inputs and flags

20 inputs
NameTypeFlagDescription
inputFILE-iOne media file. Use this or --input-folder.
input-folderFOLDER--input-folderFolder of media files to convert.
output-extensionSTRING--output-extOutput extension such as mp3, mp4, gif, or webm. Default mp4.
extra-argsSTRING--extraExtra ffmpeg arguments as one string.

Showing key inputs. ffmpeg exposes 20 inputs in total.

Full flag reference (20 inputs)
NameTypeFlagDescription
audio-bitrateSTRING--audio-bitrateAudio bitrate (e.g., 128k, 192k, 320k)
audio-codecSTRING--audio-codecAudio codec (e.g., aac, mp3, copy, libopus)
audio-filterSTRING-afAudio filter (e.g., volume=2.0, loudnorm)
crfSTRING-crfConstant Rate Factor for quality (0-51, lower is better, 18-28 is typical)
durationSTRING-tDuration of output (e.g., 00:01:00 or 60)
extra-argsSTRING--extraAdditional ffmpeg arguments (passed directly)
framerateSTRING-rFrame rate (e.g., 24, 30, 60)
output-extensionSTRING--output-extOutput file extension (e.g., mp4, mp3, gif, webm). Defaults to mp4.
output-formatSTRING-fForce output format (e.g., mp4, mp3, gif, webm)
presetSTRING-presetEncoding preset (ultrafast through veryslow)
resolutionSTRING-sOutput resolution (e.g., 1920x1080, 1280x720)
start-timeSTRING-ssStart time for trimming (e.g., 00:00:30 or 30)
subtitle-fileFILE--subtitle-fileSubtitle file to burn into video (ASS, SRT, VTT, etc.)
subtitles-folderFOLDER--subtitles-folderFolder containing subtitle files (auto-matched by filename)
video-bitrateSTRING--video-bitrateVideo bitrate (e.g., 1M, 2M, 5000k)
video-codecSTRING--video-codecVideo codec (e.g., libx264, libx265, copy, libvpx-vp9)
video-filterSTRING-vfVideo filter (e.g., scale=640:480)
cuts-fileFILE--cuts-fileFile with cuts (format: start_time end_time output_filename per line)
inputFILE-iInput file (video, audio, or image)
input-folderFOLDER--input-folderInput folder containing media files

example

Run ffmpeg

ffmpeg · command
# one file; last argv is the output folder on Trickestffmpeg -i tone.wav --output-ext mp3 /hive/out # folder of filesffmpeg --input-folder /hive/in/media --output-ext mp3 /hive/out
sample output
Processing: /hive/in/fixture/tone.wav -> /hive/out/tone.mp3Processing complete. Output in: /hive/out # tone.mp3 1413 bytes (libmp3lame, 8000 Hz mono)

guidance

Choosing ffmpeg

Use ffmpeg when you already have media and need another container or codec. Use yt-dlp when the input is still a URL. Use whisper when the next node needs speech as text.

yt-dlp

Downloads the media file. Run it first when the input is still a URL or RSS enclosure.

whisper

Speech to text from audio you already have. Convert to wav or mp3 first when the source is video.

wget

Downloads HTTP files. It does not remux or transcode.

faq

ffmpeg questions

Set -i to the input file and --output-ext to mp3. The wrapper writes stem.mp3 into the output folder.

Run ffmpeg yourself

A fixture WAV feeds ffmpeg with --output-ext mp3. The folder holds tone.mp3 from a completed run.

Facts on this page come from the live Trickest tool library.