How to Write Vocal Prompts in Flow Music AI

A useful vocal prompt tells Flow Music AI how you want the singing to feel: soft and intimate, rough-edged and conversational, or strong and open. Start with one clear direction, then add enough musical context to make it meaningful.

In the Flow Music AI generator, Custom mode includes a Styles field and a Voices category with text suggestions. These help you describe a performance. They are not a catalog of fixed singers, and a prompt does not guarantee an exact voice or delivery.

This guide shows where the words go, how the voice tags work, and how to revise a vocal brief after listening. The interface was checked on November 4, 2026. The examples below are original prompt starting points, not results from an audio comparison test.

Describe the performance before collecting tags

“Good vocals” leaves most of the decision open. A more useful brief identifies what you want the listener to notice.

Consider three parts:

Part of the brief	Question to answer	Example wording
Delivery	How should the singer perform the words?	Gentle, conversational lead vocal
Texture	What character should the voice suggest?	Lightly raspy, warm, clear
Role	How should the voice sit in the song?	One lead voice with a restrained chorus

Then connect those choices to the music: “gentle, conversational lead vocal over sparse acoustic guitar” is a clearer creative direction than an unrelated list of voice adjectives.

You do not need an instruction for every possible detail. Pick the qualities that matter to the song. If understandable words are the priority, say so and listen for that; adding five more emotional descriptions may not help you judge the result.

Put vocal direction in Styles, and song words in Lyrics

For a song with prepared lyrics:

1. Open the generator and choose Custom.
2. Keep Instrumental off.
3. Enter a working title.
4. Write the musical and vocal direction in Styles.
5. Put the words to sing in Lyrics.
6. Review the fields and displayed cost before generating.

For example, Styles could contain:

Acoustic folk, fingerpicked guitar, light percussion, soft intimate lead vocal, clear words, reflective mood, gentle lift in the chorus.

The Lyrics field should contain your song text. Putting a paragraph of sound instructions there can blur the distinction between what should be sung and how it should sound. For full lyric preparation, use the lyrics-to-song guide.

If you only have a song idea, you can include vocal direction in the Simple description instead. The Simple vs Custom guide explains which starting point fits the material you have.

What happens when you click a Voices tag?

In Custom mode, open Voices beneath Styles. The current suggestions are Male, Female, Choir, Falsetto, Raspy, Whisper, Powerful, and Soft.

Clicking a tag adds that word to the Styles text. It does not replace the whole brief or choose a saved singer. For example, clicking Soft adds “Soft”; you can then edit the field into a fuller description such as “soft lead vocal, close and conversational.”

Read the Styles field after adding tags. Clicking several suggestions can accumulate words that point in different directions. If you want one restrained lead voice, an unqualified list containing Choir, Whisper, and Powerful makes your intention harder to assess.

Some combinations can make creative sense, such as a soft verse and a stronger chorus. Describe that relationship in a sentence rather than expecting a pile of tags to define the sections for you. Treat it as a request, and check the actual performance afterward.

The currently selected V5 form allows 1,000 characters in Styles; V4 shows a 200-character allowance. Follow the counter for your selected model. A concise direction usually leaves enough room for genre and instruments as well as vocals.

Three vocal prompt starting points

These examples explore different performance goals. Adapt them to your song rather than combining all three.

Soft and close for a reflective song

Acoustic folk with fingerpicked guitar and light percussion. Soft, intimate lead vocal with clear words and a conversational delivery. Keep the verse restrained and give the chorus a gentle emotional lift.

Use this when the song's message matters more than a large, dramatic performance. During review, listen for whether quiet delivery still leaves the words understandable.

Raspy and grounded for a road-song feel

Mid-tempo roots rock with warm electric guitar, steady drums, and a lightly raspy lead vocal. Grounded, direct phrasing, a lived-in tone, and a confident chorus without constant shouting.

The useful distinction is between texture and intensity. You can ask for a rougher vocal character without asking for every line to be loud. Listen to both qualities separately.

Powerful and open for an uplifting chorus

Uplifting pop-rock with bright guitars and a steady rhythm. Clear, powerful lead vocal, measured verses, and an open, energetic chorus. Keep the main lyric easy to follow as the arrangement grows.

Here the goal is a change in energy across the song. Check whether the chorus feels larger while the words remain audible; a louder arrangement alone may not deliver the effect you intended.

Original concept illustration of descriptive vocal choices. It is not a generator screenshot, a voice library, or evidence of tested audio output.

Listen for one problem, then change one instruction

After generation, review the verse, chorus, and ending. Keep the input you used so you can compare the performance with the brief.

What you hear	What to check before another request
The vocal feels too forceful	Remove competing high-energy descriptions and request a more restrained delivery
The words are hard to follow	Review the lyric phrasing and ask for clearer delivery; do not assume a style change alone will fix it
The song includes more voices than intended	Remove Choir or group-vocal wording and describe one lead voice
There is no singing	Check Instrumental and confirm that the vocal-song inputs are complete
The chorus lacks contrast	Describe the verse-to-chorus change you want rather than adding more unrelated tags

These are revision choices, not guaranteed fixes. Another generation can change the arrangement and performance as well as the quality you were targeting. Keep a version worth saving before exploring another direction.

The generator currently displays 10 Required Credits per music-generation request. Check your balance and the live cost before submitting again. The credit-budget guide explains how to reserve attempts for purposeful revisions.

Keep the goal descriptive

A prompt can describe a vocal quality without naming a particular person. Use the qualities you need—such as gentle delivery, clear phrasing, or a rougher texture—as the creative brief. The Voices tags should not be treated as a promise of voice cloning, an identical singer across separate requests, or precise control over an existing recording.

For commercial work, review the current Terms of Service and your intended use. Once you have an acceptable version, follow the download and MP3-check guide to save and inspect the actual file.

Start with one vocal direction, put it beside the musical context in Styles, and open Flow Music AI to review the form before generating.

Sources: the live Flow Music AI generator and Terms of Service linked above. Voice suggestions, tag behavior, form limits, and displayed cost checked November 4, 2026. No paid audio generation was performed for this guide.