The open-source command-line swiss-army knife for media: decode, encode, trim, duck background music, boost voice — the engine agents reach for under every video/audio task.

Pick this when

Any local media manipulation an agent should do programmatically — it's the layer your prompts actually drive.

What makes it different

Installable by CLI, discoverable by asking the agent 'is there a package for this?'

Handles the mixing jobs (levels, fades, trims) that video skills delegate

What it does

Audio level analysis, trims, fades, format conversion, stream handling

Backs face-tracking/clipping pipelines via companion packages

What it does not do

Common wrong expectations

Doesn't have a UI — it exists to be driven by commands (or an agent)

Where it sits in a stack

Invisible infrastructure: you rarely invoke it directly; your agent does.

Shelf

Productivity & source apps

Our status · using

Standard machine utility; agents use it whenever media tasks appear.

Freshness

current

Related tools

complementsRemotioncomplementsHyperFrames

remotion: post-render audio work

hyperframes: same

Every moment the course touched it (1)

SessionCoverageMomentWhat happened
Session 14: Claude Code — Advanced Workflowsexplained2:52:49Named as the media engine for mixing, ducking, encoding; 'found it by just prompting Codex'

← All tools by category