Text-to-Music Prompting Guide
A practical reference for AI music generation tools (Suno, Udio, ElevenLabs Music).
Core Principles
1. Structured > Vague
Separate style prompts (sound) from lyrics prompts (words + structure). Don't mix them.
2. Iterate, Don't Perfect
Expect 6+ generations to land the right vibe. Treat prompts like direction to session musicians—guide, don't dictate.
3. Specific Musical Vocabulary Wins
"Punchy brass with sparkling arpeggios" beats "good instruments." Use real production terms.
Prompt Architecture
Style Prompt Formula
[Genre], [Sub-genre], [Mood], [Key instruments], [BPM], [Production style]
Examples:
Deep House, Minimal, Hypnotic, warm bass, 118 BPM, lo-fi textures
Indie Pop, Anthemic, Uplifting, atmospheric synths, emotional male vocals, 103 BPM
Ambient, Reflective, Sitar, urban meditation, lo-fi textures, Indian classical influence
Metatags for Song Structure
Use these to control arrangement:
| Tag | Purpose |
|---|---|
[Intro] | Opening section |
[Verse] | Main narrative sections |
[Pre-Chorus] | Build-up before chorus |
[Chorus] | Hook / most memorable part |
[Bridge] | Variation or mood shift |
[Breakdown] | Stripped-back section |
[Drop] | EDM energy release |
[Outro] | Closing section |
[Instrumental] | No vocals |
Example with metatags:
[Intro]
Soft piano, building anticipation
[Verse]
Melancholic, lo-fi beats, breathy female vocals
[Chorus]
Explosive, full band, anthemic, layered harmonies
[Outro]
Fade out, piano returns alone
Genre-Specific Templates
Electronic/EDM
[Genre] Techno, Progressive House, Trance
[Mood] Euphoric, Dark, Hypnotic, Energetic
[Elements] Synth pads, arpeggios, 4-on-the-floor kick, side-chain compression
[BPM] 120-140
Hip-Hop/Rap
[Genre] Boom bap, Trap, Lo-fi hip-hop
[Mood] Aggressive, Chill, Nostalgic
[Elements] 808 bass, hi-hat rolls, vinyl crackle, chopped samples
[BPM] 70-90 (trap), 85-95 (boom bap)
Pop/Indie
[Genre] Indie pop, Synth-pop, Dream pop
[Mood] Uplifting, Bittersweet, Dreamy
[Elements] Jangly guitars, atmospheric synths, layered vocals
[BPM] 100-120
Ambient/Cinematic
[Genre] Ambient, Cinematic, Neo-classical
[Mood] Ethereal, Tense, Hopeful, Melancholic
[Elements] Strings, piano, pads, field recordings, reverb
[BPM] 60-80 or free tempo
Descriptor Library
Mood/Energy
| Low Energy | Medium Energy | High Energy |
|---|---|---|
| Melancholic | Groovy | Explosive |
| Ethereal | Warm | Aggressive |
| Somber | Nostalgic | Euphoric |
| Haunting | Bittersweet | Anthemic |
| Meditative | Playful | Triumphant |
Vocal Descriptors
breathy, raspy, soulful, operatic, whispered, belted,
falsetto, husky, clear, powerful, intimate, distant,
male, female, androgynous, choir, harmonized
Production Styles
lo-fi, polished, raw, compressed, spacious, intimate,
vintage, modern, analog warmth, digital clarity,
heavy reverb, bone dry, saturated, clean
Instrument Textures
punchy, warm, bright, muddy, crisp, lush,
shimmering, gritty, smooth, aggressive, delicate
Tool-Specific Notes
Suno
- Strength: Human-like vocals, emotional nuance, complete song structures
- Best for: Full songs with clear intro/verse/chorus
- Tip: Treat vocal direction as a constraint—use consistent language across generations to build an "artist" sound
Udio
- Strength: Seamless track extension, maintaining timbre
- Best for: Extending existing clips, inpainting sections
- Tip: Better at longer instrumental passages
ElevenLabs Music
- Strength: Precise BPM and key control
- Best for: When you need exact musical specifications
- Tip: Specify key signature (e.g., "A minor") for mood control
Common Mistakes to Avoid
| ❌ Don't | ✅ Do |
|---|---|
| "Make a good song" | "Indie rock, melancholic, jangly guitars, 95 BPM" |
| Reference specific artists | Describe the style characteristics instead |
| One-shot and done | Generate 6+ variations, iterate |
| Mix style and lyrics | Keep them in separate prompt sections |
| Use generic terms | Use specific musical vocabulary |
| Expect perfection | Embrace iteration as part of the process |
Quick Reference Card
FORMULA:
[Genre], [Sub-genre], [Mood], [Instruments], [BPM], [Production]
STRUCTURE TAGS:
[Intro] [Verse] [Pre-Chorus] [Chorus] [Bridge] [Breakdown] [Drop] [Outro]
ITERATION:
Prompt → Generate 6+ → Select best → Refine prompt → Repeat
VOCAL CONSISTENCY:
Define vocal character once, use same descriptors across generations
Last updated: January 2026 Sources: howtopromptsuno.com, learnprompting.org, soundverse.ai, community discussions