Creating Music with AI: Suno, Stable Audio & Beyond

The first time I generated a song with Suno, I laughed. Not because it was bad — because it was genuinely good. I typed “lo-fi chill jazz, late-night study vibes, soft piano and brushed drums” and 30 seconds later I had a two-minute track that sounded like it belonged on a Spotify playlist.

AI music generation has crossed the threshold from “novelty” to “actually useful.” Whether you need background music for a video, a jingle for a project, or you just want to explore making music without years of training, the tools are here and they work.

Suno V5: The Full Package

Suno is the big name in AI music, and for good reason. V5 generates complete songs — vocals, instruments, arrangement, the works — from a text prompt. It understands genre, mood, tempo, and instrumentation, and it produces output that sounds remarkably polished.

How to Write a Music Prompt

The key to great Suno output is genre specificity. Vague prompts produce vague music. Specific genre tags produce tracks that sound intentional.

Genre tags that work well:

• “lo-fi chill jazz, soft piano, brushed drums, vinyl crackle”
• “90s grunge rock, distorted guitars, raw vocals, angst”
• “orchestral cinematic epic, brass fanfare, timpani, sweeping strings”
• “tropical house, steel drums, upbeat, summer vibes, 120 BPM”
• “acoustic folk ballad, fingerpicked guitar, gentle female vocals, melancholic”

Prompts that produce mush:

• “relaxing music” (too vague — relaxing how? Jazz? Ambient? Classical?)
• “a happy song” (happy covers everything from bubblegum pop to reggae)
• “something cool” (meaningless to a model)

The pattern: genre + instruments + mood + tempo. The more specific each element, the better the result. You don’t need to be a musician to write good prompts — you just need to describe what you want to hear in concrete terms.

Lyrics: Write Your Own or Let AI Generate

Suno offers two modes for lyrics. You can let the AI generate lyrics based on your prompt (it’s surprisingly decent at this), or you can write your own and paste them in.

If you write your own lyrics, keep these things in mind:

Do: Write naturally. Suno understands verse/chorus structure, rhyme, rhythm, and syllable patterns. Write like you’d sing it.

Don’t: Include stage directions in your lyrics. This is the most common mistake beginners make.

Warning: Suno sings EVERYTHING.

If your lyrics say “[Verse 1]” or “(softly)” or “*guitar solo*”, Suno will literally sing those words. It doesn’t interpret formatting markers — it vocalizes them. Keep your lyrics clean: just the words you want sung, nothing else.

Instrumental Mode

Toggle instrumental mode when you want music without vocals. This is perfect for background tracks, video soundtracks, podcast intros, and ambient music. The quality is excellent — Suno is just as good at pure instrumentals as it is at full songs with vocals.

Two Variants Per Generation

Every time you generate, Suno produces two variants of the same prompt. They’ll have the same general style but different melodies, arrangements, and interpretations. Always listen to both — sometimes the second variant is dramatically better than the first. It’s free variety built into every generation.

Extending Tracks

Suno generates tracks in segments (typically 1–2 minutes). If you want a longer track, you can extend from the end of a generated clip. The extension continues the musical ideas, key, and style of the original. This is how you build a full 3–4 minute song: generate the first section, extend, extend again.

The results are usually seamless, though occasionally the extension will shift the energy or introduce a style change. If that happens, just regenerate the extension — it’s quick.

Stable Audio 2.5: The Instrumental Specialist

Stable Audio 2.5 from Stability AI takes a different approach. It’s instrumental-only — no vocals — but it excels at producing high-quality background music and sound design.

Where it shines: Atmospheric tracks, ambient soundscapes, background music for videos, game audio, podcast beds. If you need music that supports other content without demanding attention, Stable Audio is excellent.

Where Suno wins: Full songs with vocals, lyric interpretation, genre versatility. Suno is the generalist; Stable Audio is the specialist.

Think of Stable Audio as the tool you reach for when the music needs to sit behind something else — a video, a presentation, a game. Suno is the tool when the music is the thing.

The Royalty Question

This is the question everyone asks: can you use AI-generated music commercially?

The answer, as of March 2026: yes, with caveats. Both Suno and Stable Audio grant commercial use rights on paid plans. You can use the generated tracks in YouTube videos, podcasts, games, ads, and products. You own the output.

The caveats are legal, not technical. Copyright law around AI-generated content is still evolving. No court has definitively ruled on the copyright status of AI-generated music. The practical reality is that thousands of creators are using AI music commercially every day without issues, but if you’re producing content for a major brand or broadcast, you may want to consult a lawyer.

For personal projects, social media, YouTube videos, indie games, and small business use? You’re fine. Use it.

Tips That Actually Help

Be specific about genre. “Lo-fi hip hop with jazzy piano chords, vinyl crackle, and a slow boom-bap drum pattern” will always produce better results than “chill beats.”

Mention specific instruments. “Acoustic guitar, upright bass, brushed snare” gives the model clear targets. “Nice instruments” gives it nothing.

Set the mood with adjectives. “Melancholic,” “triumphant,” “eerie,” “playful,” “aggressive” — these words carry enormous weight in how the model shapes the track.

Include tempo when it matters. “120 BPM” or “slow tempo” or “fast-paced” helps the model match the energy you need, especially for video sync.

Listen to both variants. Always. The difference between variant 1 and variant 2 can be the difference between “meh” and “perfect.”

The secret nobody mentions: AI music generation is at its best when you know what you want but lack the technical ability to produce it. If you can hear the song in your head — the genre, the instruments, the mood — but you can’t play guitar or don’t own a studio, these tools bridge that gap. The people getting the best results aren’t musicians. They’re music listeners who know exactly what they like.

Both Suno and Stable Audio are available on Zubnet with per-generation pricing — no monthly subscription to a music platform you only use occasionally. Generate when you need it, pay for what you use.