Game devlog video with no voice-over: captures, a bed, cue cards

To make a game devlog video without recording a voice-over, trim a few gameplay captures, render them on Timeline 1.0 with audio mode silence plus a music bed as the soundtrack, then burn each feature name as a caption cue. At the listed rates a 45-second devlog with five trims, one Music Router track, one timeline render and one caption job costs $0.525. Nothing in the chain needs a microphone.

Why a silent spine and a bed?

Timeline 1.0 normally takes an audio spine, usually a voice-over, and lays video slots on top of it. The docs also define audio.mode set to silence: you declare the length (1 to 1800 seconds) and send no audio file. The soundtrack field adds an optional bed with a url, loop and fade-out, so the devlog gets music with no narration. Only the duck setting needs a real spine, because there is nothing to duck under in silence mode.

The captured gameplay audio is not part of that render. If you want the game's sound, make it the spine instead: detach the audio from the capture with audio detach ($0.01) and pass that file as audio.url.

What does the render request look like?

Clips are placed by declared start times. The first slot must start at 0, later starts must increase, and a transition belongs only on slots after the first (fade, wipeleft, wiperight, slideup, slidedown or dissolve, up to 1 second). The default output is 1080 by 1920, so set width and height for a landscape devlog. The sample shows three of the five slots.

{
  "audio": { "mode": "silence", "duration_seconds": 45 },
  "video": [
    { "source_url": "https://media.sume.com/artifacts/artf_demo/combat.mp4", "start": 0, "duration": 9 },
    { "source_url": "https://media.sume.com/artifacts/artf_demo/boss.mp4", "start": 9, "duration": 9,
      "transition": { "type": "fade", "duration": 0.25 } },
    { "source_url": "https://media.sume.com/artifacts/artf_demo/map.mp4", "start": 18, "duration": 27,
      "transition": { "type": "wipeleft", "duration": 0.5 } }
  ],
  "soundtrack": {
    "url": "https://media.sume.com/artifacts/artf_demo/bed.mp3",
    "loop": true,
    "fade_out_seconds": 3
  },
  "output": { "width": 1920, "height": 1080 }
}

What does each step cost?

Run the unbilled plan route first. It returns duration, segment count and estimated cost without creating a job, so you see the minute count before paying.

How do the feature cards get on screen?

The timeline compiler does not draw text. Text goes on in the last step: send the finished render to video captions with cues, one per feature, such as New dodge roll at 2 seconds for 4 seconds. The captions page says cues need no speech-to-text, which matters because this render has no speech to find.

Keep the bed's prompt specific about length and mood, since the Music Router rejects a duration field and takes length from the prompt. If the bed is shorter than the video, set loop to true or the render warns that the bed ends early.