Use Cases · 7 min read

How do you make podcast music with AI that doesn't fight the voice?

Published 23 March 2026

Short answer

Make a small kit in one style: an intro, an outro, a few short stings and an instrumental bed with no lead melody. In Autunes, use Instrumental only for beds and Short-Form for 15 to 60 second pieces. Under speech, keep the bed well below the voice and dip it further while people talk. Use a paid plan or the ₹9 licence per track for a published show.

Most podcast music problems are not about taste. They are about the music competing with the voice. A beautiful track with a strong melody and vocals becomes tiring under twenty minutes of conversation, and listeners cannot say why they switched off.

This guide covers the full sound kit: intro, outro, transition stings and the music bed that sits under speech, plus the loudness and ducking basics that make it all work. If you only need a short intro, start there and come back for the rest.

The parts of a podcast sound kit

A complete kit is small. Most shows need only four or five pieces.

  • Intro: 5 to 30 seconds, the show's signature.
  • Outro: 20 to 60 seconds, calmer, with room for the host to thank guests and give the call to action over it.
  • Transition stings: 2 to 5 seconds, to move between segments.
  • Music bed: an instrumental that sits under talk, for openings, recaps or narrated sections.
  • Optional: a sponsor or ad-break bumper, and a theme for a recurring segment.

What makes a good bed under speech

A bed should support the voice without asking for attention. The human voice sits mostly in the middle of the frequency range, so music that fills the same space with a busy melody gets in the way.

  • No vocals, ever. Words in the music compete with words in the conversation.
  • No strong lead melody; chords, pads, soft keys, light plucks and gentle rhythm work better.
  • A steady, moderate tempo, often around 80 to 110 BPM, with few dramatic changes.
  • Simple and loop-friendly, so it can run for a minute or be repeated.
  • The same instruments and mood as your intro, so the kit sounds like one family.

Making the kit as one family

Write one base style prompt and reuse it for each piece, changing only the job. For example, a base of "warm lo-fi, soft electric piano, light brushed drums, mellow bass" becomes:

  • Intro: "warm lo-fi, soft electric piano, light brushed drums, mellow bass, inviting podcast intro, clean ending"
  • Outro: "same style, slower and calmer, gentle fade feel, room for speech"
  • Bed: "same style, minimal, no lead melody, steady, instrumental, background under speech"
  • Sting: "same style, short two-bar flourish, clean ending"

Step by step in Autunes

Use the tool that fits each piece.

  • Intro and outro: in the Short-Form tab, the Podcast Intro / Outro template makes instrumental clips of 15 or 30 seconds for 2 or 3 credits.
  • Bed: in the Song tab, turn on Instrumental only and use the bed prompt. Songs run up to 3 minutes on Free and Pro and 5 on Creator. For a 60-second loopable piece, Short-Form's Hold / Waiting Music and Game Background templates are built to loop.
  • Stings: cut 2 to 5 seconds from the end of an intro or outro at a natural phrase end.
  • Alternative takes: a variation costs 5 credits and keeps the same brief.
  • On Pro and Creator, split a track into 6 stems for 5 credits and build a bed from just the piano, bass and drums, leaving out the melody.

Loudness and ducking basics

Ducking means turning the music down while someone speaks and back up in the gaps. Any editor can do it, either by drawing volume changes by hand or with a sidechain compressor that reacts to the voice. Keep it smooth: fades of around half a second to a second sound natural, while instant jumps sound like a mistake.

How low should the bed go? There is no single number, but under speech it usually sits roughly 15 to 25 dB below the voice. If you notice the music while listening to the words, it is too loud. Check on phone speakers and cheap earphones as well as good headphones.

For the finished episode, a widely used target for stereo podcasts is around -16 LUFS integrated loudness, with true peaks no higher than about -1 dB. Many apps adjust loudness on playback anyway, but a consistent level means your music and voices stay balanced across episodes. A gentle cut in the music around the middle frequencies, where speech sits, also helps voices stand out.

Narrated shows and AudPod

If your show is narrated rather than a conversation, such as stories, kathas, history or explainers, AudPod does the narration and the background music together. It works from an idea or your own script, in 42 languages, with a male or female narrator. 3 or 5 minutes cost 10 credits; Pro allows 10 minutes and Creator 20. Your first AudPod is free, up to 5 minutes.

Licensing for a published show

Free-plan tracks are for personal use. For a published podcast on any platform, especially one with sponsors, use a paid plan or buy the ₹9 single-song licence for each track (US$1.79 or €1.59 outside India). A full kit is four or five tracks, so Pro, at ₹599 for 30 days with 300 credits, is often simpler. There is no royalty and no credit to Autunes required.

Common mistakes

Most of these are mixing problems rather than music problems.

  • A bed with vocals or a strong melody under conversation.
  • Music at the same level under speech as in the intro.
  • Abrupt ducking that pumps up and down between every sentence.
  • Pieces in different styles, so the show sounds stitched together.
  • Checking the mix only on studio headphones.

Try it yourself

Make a full song from text — free to start, no skills needed.

Start creating free

Frequently asked questions

How loud should background music be under a podcast voice?

Low enough that you do not notice it while following the words, often roughly 15 to 25 dB below the voice. Trust your ears on phone speakers too.

What loudness should a podcast be mastered to?

A widely used target for stereo podcasts is around -16 LUFS integrated, with true peaks no higher than about -1 dB.

Should podcast background music have vocals?

No. Sung words compete with spoken words. Use instrumentals with no strong lead melody.

How do I make the music loop under a long segment?

Use a loop-friendly 60-second Short-Form template, or a longer instrumental song, and crossfade repeats in your editor.

Can I take the melody out of a track I already like?

On Pro and Creator, split it into stems and keep only the parts that suit a bed, such as piano, bass and drums.

How many credits does a full podcast kit cost?

An intro and outro from Short-Form cost 2 or 3 credits each, and an instrumental bed 10 or 15. A whole kit with a few variations fits easily within Pro's 300 credits.

Related reads