How to Use TTSMP3

Enter Your Text

Start by typing or pasting your text into the main editor. TTSMP3 supports up to 5,000 characters per generation.

You can also:

• Upload a TXT file: Click “📁 Upload TXT” to import text from a file
• Paste from anywhere: Copy from Word, Google Docs, or any text editor
• Use emojis: They’ll be automatically filtered out during processing

Choose Your Voice

Select from 15+ professional AI voices in the right panel. Each voice has unique characteristics:

• American Female: Heart, Alloy, Bella, Sarah, Sky, Nicole, River
• American Male: Adam, Echo, Michael, Ryan
• British Female: Emma, Jessica
• British Male: George, Lewis

Adjust Settings (Optional)

Fine-tune your audio output with these settings:

• Output Format:
  • • WAV: Lossless quality, larger file size (~10MB per minute)
  • • MP3: Compressed, smaller file size (~1MB per minute)
• Playback Speed: Adjust from 0.5x (slow) to 1.5x (fast) without affecting pitch

Generate Audio

Click the “🎵 Generate Audio” button. Here’s what happens:

• First Generation (10-20 seconds): The 82M-parameter AI model downloads and loads into your browser
• Subsequent Generations (2-3 seconds): Instant processing using the cached model
• Long Texts: Automatically split into chunks and seamlessly stitched together

You’ll see a progress bar showing:

• Model loading status
• Processing chunks (e.g., “3 / 8” means chunk 3 of 8)
• Final assembly and crossfade application

Download Your Audio

Once generation is complete:

• Preview: Use the built-in audio player to listen before downloading
• Download: Click “⬇️ Download Audio” to save your MP3/WAV file
• Filename: Files are named ttsmp3_[timestamp].[format]

Using Director Mode

Create dynamic conversations with multiple AI voices using simple syntax:

Director Mode Rules:

• Start each line with [VoiceName]:
• Voice names must match exactly: Heart, Adam, Emma, Michael, etc.
• Each voice segment can be multiple sentences
• Leave blank lines between speakers for natural pauses