Prompting Guide
Master the art of writing effective prompts to get the best results from Wubble AI
What you'll learn
Introduction
Wubble's AI understands natural language and can generate music, voiceovers, and sound effects from simple text descriptions. The quality of your prompts directly impacts the quality of the results. This guide will teach you how to communicate effectively with Wubble AI to get exactly what you need.
Why Good Prompting Matters
The more specific and descriptive your prompts, the better Wubble AI can understand your creative vision. Think of prompting as giving clear directions to a talented collaborator who wants to help bring your ideas to life.
General Prompting Principles
These core principles apply to all types of audio generation in Wubble:
1. Be Specific and Descriptive
Include concrete details about style, mood, instruments, tempo, and intended use. Avoid vague descriptions.
2. Provide Context and Use Case
Mention what the audio will be used for (e.g., YouTube video, podcast intro, game background). This helps the AI understand the right tone and duration.
3. Use Natural Language
Write prompts as if you're talking to a creative collaborator. You don't need special syntax or technical jargon (unless you want specific technical results).
4. Iterate and Refine
Start with a clear prompt, review the results, and refine by asking for adjustments. Wubble remembers context within a conversation.
Music Generation Prompting
Music generation works best when you describe musical characteristics, mood, and intended use. Wubble can create anything from short background loops to full tracks with vocals.
What You Can Specify
Genre & Style
Electronic, jazz, rock, classical, hip-hop, lo-fi, cinematic, etc.
Mood & Energy
Upbeat, calm, energetic, melancholic, dramatic (low, medium, or high energy)
Instruments
Piano, guitar, synths, drums, bass, strings, brass, etc.
Duration
3 to 300 seconds (default: 30s)
Theme/Use Case
Gaming, Corporate, Vlogs, Cinematic, Sports, Technology, Documentary
Vocals & Lyrics
Instrumental or with vocals (provide lyrics or let AI generate them)
Character Limits & Features
Prompt length: 1-4,100 characters
Lyrics length: Maximum 3,000 characters
Batch variations: Automatically generates 2-3 variations per request
Playlist generation: Create 2-10 tracks in a single request
Seamless loops: Perfect for game background music
Good Examples
"Create upbeat electronic music with driving synth arpeggios and strong bass. Make it energetic and modern, perfect for a tech product demo video. Duration: 60 seconds."
"Smooth jazz track with gentle piano melodies, soft upright bass, and light brush drums. Calm and sophisticated mood, perfect for a café background. 2 minutes long."
Bad Examples
"Make some music for my video"
Problem: No style, mood, duration, or context provided
"A 1-minute song that fits the main brand identity, with elements of Pop, Jazz, and Hip Hop"
Problem: References abstract concepts instead of concrete musical characteristics
Pro Tip
For songs with vocals, you can either provide your own lyrics or ask Wubble to generate them. Just mention "generate lyrics" or describe what the song should be about. Don't leave lyrics empty for vocal tracks.
Voice Generation Prompting
Voice generation creates realistic AI voiceovers from your script. You can specify voice characteristics, language, emotional tone, and use special formatting for expressive delivery.
Voice Characteristics
Gender
Male, female, neutral, non-binary
Accent
American, British, Australian, Standard
Age
Young, middle-aged
Descriptive Traits
Classy, professional, sassy, hyped, mature, calm, rough, confident
Use Case
Conversational, social media, narrative, educational, entertainment, characters
Language
30+ languages supported (EN, ES, FR, DE, JA, ZH, etc.)
Character Limits & Features
Script length: 1-40,000 characters
Expressive formatting: Only works for scripts ≤5,000 characters
Stability control: 0.0-1.0 for voice consistency (higher = more consistent)
Speaking rate: 0.7-1.2 speed multiplier (1.0 = normal)
Formatting Styles
Expressive Short (≤5,000 characters)
Use bracket tags to add emotional expression and vocal effects:
Narrative Long (>5,000 characters)
For longer scripts, use narrative descriptions or SSML tags. Bracket tags won't work here.
Neutral
Plain text for straightforward delivery without special formatting.
Emphasis Markers
Use asterisks around words to add emphasis:
"This is *absolutely* incredible! You *must* try this feature."
Good Examples
Prompt: "Young American female voice, professional and hyped, for social media"
Script: "[excited] Hey everyone! Today I'm showing you the *coolest* new app that's going to change the way you work. [laughs] Trust me, you need to see this!"
Prompt: "Middle-aged British male voice, calm and professional, for documentary narration"
Script: "The ancient civilization thrived for over three thousand years, leaving behind monuments that continue to inspire us today."
Important
If you specify a non-English language, make sure your script is written in that language. The AI doesn't automatically translate. For example, if you select Spanish (ES), write your script in Spanish.
Sound Effects Generation Prompting
Sound effects (SFX) generation requires precise, concise descriptions. Think about the specific sound you need and describe it clearly.
Character Limits & Features
Prompt length: 1-500 characters (be concise!)
Duration: 0.5-30 seconds (default: 5.0s)
Batch variations: Automatically generates 2-3 variations per request
Speed tip: Shorter durations generate faster
SFX Prompting Tips
Be Precise
Describe the exact sound you need: "thunder crash", "door creak", "footsteps on gravel"
Include Context
Mention the context or purpose: "UI button click", "game level complete notification", "transition swoosh"
Describe Characteristics
Add adjectives: "soft", "loud", "metallic", "wooden", "digital", "organic"
Examples by Category
UI Sounds
• "Soft button click, digital, pleasant"
• "Success notification chime, uplifting"
• "Error beep, attention-grabbing"
Ambient & Background
• "Forest ambience with birds chirping and light breeze"
• "City street noise, distant traffic, urban atmosphere"
• "Calm ocean waves on a beach"
Impacts & Actions
• "Heavy door slam, wooden, echoing"
• "Footsteps on hardwood floor, steady pace"
• "Glass breaking, sharp, dramatic"
Cinematic & Transitions
• "Epic cinematic braam, horror movie impact sound"
• "Smooth whoosh transition, futuristic"
• "Rising tension drone, ominous"
Pro Tip
With the 500-character limit, focus on the most important characteristics. Don't write paragraph descriptions—be clear and concise. You'll get 2-3 variations automatically, so you can pick the best one.
Multi-Asset Generation
You can request multiple types of audio assets in a single prompt. Wubble's planner agent intelligently breaks down your request and generates each asset type using the appropriate tools.
How It Works
When you ask for multiple assets, the planner creates separate generation steps that run in parallel when possible:
"Create upbeat background music for a podcast intro, a voiceover saying 'Welcome to Tech Talk' in a professional voice, and a notification sound for the intro jingle."
The planner will automatically create three parallel tasks: one for music, one for voiceover, and one for SFX.
Best Practices
Be clear about which asset type each part of your request refers to
Provide specific details for each asset as you would in individual prompts
You can request all three types (music, voice, SFX) or any combination
The AI will organize and track each asset separately in your project
Advanced Prompting Tips
Take your prompting to the next level with these advanced techniques:
Use Precise Musical Terminology
Reference specific instruments, chord progressions, time signatures, or BPM for expert-level control: "120 BPM, 4/4 time, C major with ii-V-I progression"
Layer Descriptors
Combine multiple descriptors for nuanced results: "melancholic yet hopeful, nostalgic with a modern twist"
Request Seamless Loops
For game background music, specify: "Create a seamless loop that can repeat infinitely without jarring transitions"
Generate Playlists
Create 2-10 related tracks at once: "Generate a playlist of 5 lo-fi hip-hop tracks for studying, each 2 minutes long, with variations in mood"
Control Voice Consistency
For character voices or brand consistency, request higher stability: "Use high stability for a consistent character voice throughout the narration"
Adjust Speaking Pace
Control pacing for dramatic effect: "Slightly slower speaking rate for dramatic documentary narration" or "Fast-paced for energetic ad read"
Combine Expression Techniques
Mix bracket tags with emphasis markers: "[excited] This is *absolutely* amazing! [whispers] You won't believe what happens next."
Understand Automatic Variations
Every music and SFX request generates 2-3 variations automatically. Review all variations to find the perfect match for your project.
Common Mistakes to Avoid
Avoid these common pitfalls to get better results faster:
Being Too Vague or Generic
"Make music" or "Create a voiceover" without details
Exceeding Character Limits
Music prompts: 4,100 max | Voice scripts: 40,000 max | Lyrics: 3,000 max | SFX prompts: 500 max
Using Meta-Information Instead of Descriptions
Describing "brand identity" or "corporate feel" instead of actual musical characteristics
Wrong Language for Voice Scripts
Selecting Spanish but writing script in English (scripts aren't auto-translated)
Confusing Asset Types
Using music prompts for SFX or vice versa (they have different parameters and optimizations)
Leaving Lyrics Empty for Vocal Tracks
If you want a song with vocals, either provide lyrics or ask Wubble to generate them
Using Expressive Tags on Long Scripts
Bracket tags like [excited] only work on scripts ≤5,000 characters
Not Using Emphasis Markers
Missing opportunities to add clarity with *word* emphasis in voice scripts
Quick Reference
| Asset Type | Character Limit | Duration Range | Key Features |
|---|---|---|---|
| Music | Prompt: 1-4,100 Lyrics: 0-3,000 | 3-300 seconds | Playlists, loops, variations |
| Voice | Script: 1-40,000 | Based on script length | 30+ languages, bracket tags |
| SFX | Prompt: 1-500 | 0.5-30 seconds | Fast generation, variations |
Next Steps
Now that you understand how to write effective prompts, explore detailed guides for each capability:
You're Ready to Create!
You now have all the knowledge you need to write effective prompts for Wubble AI. Start experimenting with different prompting techniques and discover what works best for your creative projects.