Music for narration and voice-over

A voice carries the video, and the music sits underneath. Make a bed with no lead melody in the way, steady from the first word to the last.

1 version with stems · instrumental

Use in ChatGPT
Try

This is an example. Describe your own above and it comes like this too: solo a part, redo one, download them, or edit it in Studio.

Songs made here

Made from these words on this page, each with its stems. Move a level, or press M or S, to hear a part on its own.

Scary story. a dark storytelling bed for a scary story channel, a low drone, slow heartbeat kick and distant piano notes

How it works

  1. Start from the words in the box. "a soft, warm background loop to talk over, brushed drums, round bass and quiet electric piano, no lead melody". Change any of them, or pick an idea under the box.
  2. Pick the shape. Seamless loop is on, for music that repeats without a click. Pick a tempo, a key and a length, or leave them on Auto.
  3. Get your track with its stems. Play it, download the mix or the stems, or open it in Studio to change the parts.

A bed, not a song

Music under a voice has one job: to be there without being heard over the words. That changes what to ask for.

  • No lead melody. Say it in the description. A melody up front competes with the voice for the same attention.
  • Low and soft sounds: pads, brushed drums, a round bass, quiet electric piano. A voice sits in the middle of the range, so a bed built below and around it keeps every word clear.
  • Steady from start to end. Ask for the same energy all the way through, so the music never swells in the middle of a sentence.

Turn on Seamless loop and the bed runs as long as the recording, with no gap where it starts again.

Storytelling, scary stories and explainers

Story channels need a mood that holds for ten minutes or more. For a mystery or true story, ask for a curious pulse and plucked strings. For a scary story, a low drone, a slow heartbeat kick and the odd distant piano note, with long quiet gaps so the voice does the scaring. For an explainer or a recorded slide deck, keep it calm and even: soft pads, warm bass and a gentle beat, nothing that pulls the eye off the slide.

Every track comes as five stems (drums, bass, guitar, keys and other). Mute the keys while someone talks and bring them back between chapters, or drop the drums for the quietest part of the story.

Mixing it under the voice

Set the music far below the voice, about 20 dB under it while someone speaks, and let it come up in the pauses. Most editors can duck it for you whenever the voice track has sound. MP3 downloads are unlimited on every plan, Free included, so make a few beds and keep the one you can talk over for an hour.

Questions

What tempo works under a voice?

Slow to medium, around 70 to 95 BPM. Faster beds make a voice sound rushed, and very slow ones can drag. Pick a tempo in the box or name one in the words.

Can I use it for a podcast or an audiobook?

Yes. Every track is royalty-free for commercial use, so it can sit under a podcast episode, an audiobook chapter, a course or a YouTube video.

How do I keep the music from fighting my voice?

Ask for no lead melody, and name low and soft sounds: pads, brushed drums, a round bass, quiet piano. A voice lives in the middle of the range, so a bed built below and around it leaves the words clear.

Can I use the music in a commercial game or video?

Yes. Every song is royalty-free on every plan, Free included: use it in commercial games, videos, streams, podcasts and apps.

Can I use it in Claude or ChatGPT?

Yes. Add https://gen.daydream.live/mcp as a custom connector in Claude, or as a custom MCP server in ChatGPT, or add the Daydream app from ChatGPT's directory, and sign in with your Daydream account. Then describe a song in the chat: it plays there with its stems in their own lanes, and if you ask for new drums it makes that part again with the other four as they are. It spends the same credits as this page, 5 a song. How to add it.