📖 Music Generation Guide

Getting Started

VelaSong uses advanced AI music and vocal generation models to create original songs from text descriptions and lyrics. Here's how to get the best results:

What VelaSong Generates

  • Fully produced songs with vocals
  • Instrumental tracks
  • Songs in dozens of genres
  • Custom lyrics (optional)
  • Multiple variations of the same idea

1. Write a Good Prompt

Describe the mood, style, and instruments you want. Example: "Upbeat pop song with acoustic guitar and synth pads, happy summer vibes"

Prompt Templates

Pop: "Upbeat pop track with female vocals, acoustic guitar, and warm summer energy."

Hip-Hop: "Laid-back hip-hop beat with deep bass, mellow keys, and confident flow."

Cinematic: "Emotional orchestral score with strings, piano, and slow dramatic build."

2. Add Lyrics (Optional)

Use [verse], [chorus], [bridge] tags to structure your lyrics. Leave blank for instrumental tracks.

  • Keep lines short and singable
  • Avoid overly complex rhyme schemes
  • Use emotional or descriptive language
  • You can leave sections blank for instrumental parts
[verse]
Walking down the city streets
Thinking 'bout the way we used to be

[chorus]
I still feel you in the rhythm
Every beat reminds me of you

Common Mistakes to Avoid

  • Prompts that are too short (for example: "happy song")
  • Conflicting instructions (for example: "slow fast upbeat sad")
  • Overly long lyrics
  • Missing section tags like [verse], [chorus], [bridge]

Best Practices

  • Be specific about mood and instruments
  • Choose a genre that matches your idea
  • Start with shorter durations to preview results quickly
  • Refine your prompt after each generation

3. Choose Genre & Duration

Select a genre that matches your vision. Shorter durations (15-30s) are good for previews, while longer ones (60-120s) create full sections.

🎁 Free Trial

New accounts receive a free trial song — no credit card required. Try VelaSong risk-free and hear the results for yourself.

4. Credit Costs

Each generation costs credits based on duration. Free accounts can generate 2 trial songs and 3 AI lyrics before purchasing credits.

FAQ

Can I generate songs without lyrics?
Yes. You can leave lyrics blank and generate an instrumental direction in your prompt.

Do I own the music I create?
Yes. Any music you generate with VelaSong is fully owned by you. You're free to download it, use it, share it, or publish it however you like. If you choose to make a song public in the Gallery, you simply grant VelaSong permission to display it on the platform, but ownership stays with you.

Why did my song sound different than expected?
AI generation is probabilistic. Tighten your prompt, simplify instructions, and regenerate for better alignment.

How does the free trial work?
Sign up and you can generate your first song completely free. No credit card needed.

How many credits does each duration use?
Credit usage scales with duration. Check the current conversion on the Credits page before generating.

🎛️ Advanced Tips (click to expand)

Vocal Control Tags

Place these in your lyrics to control how vocals are performed in specific sections:

[raspy vocal]
[whispered]
[falsetto]
[powerful belting]
[spoken word]
[harmonies]
[ad-lib]

Energy & Emotion Tags

Control the intensity and emotional feel of each section:

[high energy]
[low energy]
[building energy]
[explosive]
[melancholic]
[euphoric]
[dreamy]
[aggressive]

Extended Structure Tags

Beyond [verse], [chorus], and [bridge], you can use these for finer control over song structure:

[Intro]
[Pre-Chorus]
[Post-Chorus]
[Outro]
[Instrumental]
[Guitar Solo]
[Piano Interlude]
[Build]
[Drop]
[Breakdown]
[Fade Out]

Combining Tags

Add a modifier after a dash to give sections more character. Keep it to one modifier — stacking too many confuses the model.

[Chorus - anthemic]
[Bridge - whispered]
[Verse - melancholic]
[Outro - fade out]

Prompt Power Techniques

  • UPPERCASE text signals higher vocal intensity (e.g., "WE ARE THE CHAMPIONS!")
  • Parentheses create background vocals: "We rise together (together)"
  • Combine dimensions for precision: style + emotion + instruments + timbre (e.g., "80s synthwave, nostalgic, analog synth pads, warm and dreamy")

Ready to create your first song?

🎵 Start Generating