Corporate onboarding video music: a 5-minute ducked bed

For a 5-minute onboarding video, generate one neutral, upbeat instrumental of about five minutes and mix it under the narration with duck_db in Timeline. On Sume that is one generation at $0.125 and a five-minute render at $0.50, so $0.625 for the audio and assembly.

Onboarding content is long and informational. A bed that is too lively tires the viewer, so the brief aims at quiet optimism with very little movement and no vocals. A long piece also benefits from named sections so the music has some shape without a distracting climax.

The brief to send

The sections below give a soft intro, a steady middle, and a light close, each with its own small change.

curl -X POST https://api.sume.com/v1/music-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: corporate-onboarding-video-music-5-minut-001" \
  -d '{
      "model": "sume/music-auto",
      "prompt": "Friendly corporate onboarding bed, 100 BPM, C major. Soft electric piano, light acoustic guitar, gentle bass, minimal soft percussion. [0:00-0:30] Intro. [0:30-4:30] Steady groove with slow variation. [4:30-5:00] Close. Optimistic and neutral, clean modern production. Instrumental, no vocals. A 5-minute track."
    }'

Why each axis is set that way

Getting the length right

Ask for five minutes in the prompt and verify the real length. With the narration as the spine, set soundtrack.duck_db so the bed dips while someone speaks; ducking is rejected on a silent spine. Use soundtrack.gain_db for the overall bed level and fade_out_seconds for the close.

If the narration runs longer than the track, soundtrack.loop repeats it. A five-minute chapter is easier to produce than a 20-minute video, so split a long course into chapters and render each separately.

Before paying for a render, call POST /v1/timeline-1.0/plan with the same body. It is unbilled and reports the billable minutes, which for this 300-second video is 5. Then send the real POST /v1/timeline-1.0/render with an Idempotency-Key, wait up to 30 seconds in sync mode, and poll GET /v1/jobs/:id/status if it is still queued or processing. Statuses are queued, processing, completed, failed and canceled, and a 429 queue_full means try again shortly.

What it costs

A five-minute render is five started minutes, $0.50. Splitting the video into five one-minute modules would cost the same $0.50 in render, but five generations, $0.625 for music instead of $0.125, so one bed per video is cheaper than one bed per module. Scaling is simple arithmetic. One finished video here is $0.625 of audio and assembly, so ten of them are $6.25 and fifty are $31.25, before any retries. Failed jobs are not billed as accepted generations, but a regeneration you ask for is a new job, so budget one or two extra takes for a hero clip.

Check before you publish

Use one brand bed across the onboarding series. Because generation is flat-priced, you could also give each department its own variant for the cost of $0.125 each.

Keep the prompt in your style guide so new videos sound consistent.

The hand-off between the two calls is the step people miss. When the music job completes, GET /v1/jobs/:id/result returns result.artifacts[] with an audio/mpeg file on media.sume.com. Timeline only accepts workspace media URLs, so import any outside file through POST /v1/media-imports first, and send the resulting URL as soundtrack.url. The Music docs also note that job.request.routed_model names the engine that actually ran, which is worth saving with the file.

• Do not use music to carry policy information; the narration does that.
• Use captions for compliance content and keep the bed low.
• Check the final level on laptop speakers, which is how most people watch.
• Confirm usage terms for generated music with Sume before internal or external distribution.