[Feature]: Import and export a voiceover through MCP #1280

Search existing issues

• I have searched the existing issues

Is your feature request related to a problem?

I drive OpenScreen from an MCP client to make a narrated product demo. Recording, timeline edits, and MP4/GIF export are already there. The narration is not a first-class file I can hand in and take back out.

I can produce the voiceover outside OpenScreen, or record one in the editor, but MCP cannot place that narration on the project as the voiceover, and it cannot write the project's voiceover back out as its own audio file. The only finished artifact is a flattened video, so replacing or reusing the narration means another tool and another encode.

Describe the solution you'd like

Add a voiceover round-trip on the MCP server, with the same operations on the CLI:

Import

Accept a narration file (mp3, wav, or m4a) and place it on the timeline as the voiceover, with a start offset.
Choose whether it replaces the take's microphone and system audio, or mixes with them.
Keep it as an editable track (trim, fade, gain), not only a file sitting in the media library.
Export

Write that voiceover track out as a standalone audio file (wav or m4a), using the timeline's offset, trims, and gain.
Optionally also write the mixed program audio.
Include this voiceover in exportVideo's MP4, the same way the editor's export includes an imported audio track.

Describe alternatives you've considered

openscreen export --audio mixes a file into the MP4 after the render. It does not add the file to the project, and it does not give the narration back as audio.
importMedia can add an audio file to the media library, and addAudio can lay a track. There is still no way to export that track as audio, and exportVideo only takes the Export dialog's picture options.
Mixing afterwards in another tool is the extra step this request is meant to remove.

Additional context

The editor can already record a voiceover (shortcut V) and add audio from the timeline. #350 asked for voiceover and background music inside the editor; that issue is closed. This request is the MCP and CLI round-trip: bring a narration in, and take that narration back out as audio. I searched the existing issues before opening this.