AI jingle generator from text: make a podcast intro sting

An AI jingle generator turns a written description into a short piece of music: you describe the mood, tempo, instruments, and length, and a music model returns an audio file. Sume's Music Router has no length setting, so ask for the length in the prompt, then cut the finished track to the exact second with a Timeline audio split. The request has no lyrics field, so write the brief for an instrumental sting.

The steps come from Sume's Music Router, Music 1.0, and Timeline audio docs, read on 2026-09-27, and prices from the code behind API pricing. Prompt craft in general is in AI music prompt examples.

How do I write a prompt for a jingle or podcast intro?

Write the sting as a music brief: the emotion, the genre, a tempo in BPM, the key, two to four instruments, one named moment, and the era or production, then the length and “Instrumental, no vocals.” Name where the ending should land; that is where you will look for the cut.

• Podcast intro: “Curious, upbeat jazz-hop intro sting, 94 BPM, B-flat major. Walking upright bass, a dusty drum break, a muted trumpet hook. Two bars of drums, the trumpet hook enters at 0:03, and the band lands on one held chord at 0:08. Warm and close, 2010s production. An 8-second track. Instrumental, no vocals.”
• Sonic logo for an ad: “Bright, playful five-note mallet motif, 128 BPM, C major. Marimba, glockenspiel, a soft sub hit under the last note. One rising phrase that ends on a held chord at 0:04. Modern and crisp. A 5-second track. Instrumental, no vocals.”
• Outro bed under a sign-off: “Relaxed, friendly lo-fi outro, 86 BPM, F major. Dusty Rhodes, soft brushed drums, round bass. A steady groove that thins out after 0:20. Late-2010s bedroom production. A 30-second track. Instrumental, no vocals, no spoken word.” Add “no spoken word” only when a voice will sit on top.

Can I set a jingle's exact length?

Not in the music request. duration and duration_seconds are rejected, and length is steered only by the prompt (“an 8-second track”, or section markers such as [0:00-0:08]). The Music 1.0 docs describe the output as full-length structured songs up to a few minutes, and they call a brief's details creative directions, not guaranteed output settings. So treat the length you ask for as a target, and make the cut the step that sets it.

How do I cut the jingle to an exact length?

Send the track's URL to POST /v1/timeline-1.0/audio with operation: "split" and the range you want. The music result is already a media.sume.com file in your workspace, which is what Timeline audio requires. Send an Idempotency-Key; it is required.

• ranges takes 1–20 { start, end? } ranges in seconds, and ranges may overlap, so one job can cut an 8-second intro and a 5-second version from the same track.
• Each segment in the result has its own audio_url. There is no GET /v1/timeline-1.0/audio/:id: poll GET /v1/jobs/:id/status and GET /v1/jobs/:id/result.
• output.format is wav by default (pcm_s16le, sample-exact) or mp3, which is smaller but re-adds priming padding at every edge. Keep wav if the file will be joined again.
• Timeline audio documents no fade, and in current code a split is a straight cut, so pick an end point after the music has resolved.

curl -X POST https://api.sume.com/v1/timeline-1.0/audio \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: intro-sting-cut-001" \
  -d '{
    "operation": "split",
    "url": "https://media.sume.com/artifacts/artf_demo/intro-sting.mp3",
    "ranges": [{ "start": 0, "end": 8 }, { "start": 0, "end": 5 }]
  }'

Can the jingle say or sing my brand name?

Not as a documented feature. The music request has no lyrics field, the docs' example briefs are all instrumental, and result.lyrics is model-reported metadata, not an input. For a spoken name, make a voice line with text to speech. Timeline audio concat can join that line and the sting into one file only when both share one channel layout; otherwise the job fails with audio_parts_channel_mismatch.

How much does an AI jingle cost?

Two jobs: one generation and one split, each plus a 5.5% agent fee by default. One split job can return several cuts.