​
neo.one model

Our fast, efficient TTS model, designed for quick processing and resource optimization across a wide range of languages and use cases. View all neo.one speakers
  1. Content Localization: Quickly translate and voice-over content for multiple languages.
  2. News Broadcasting: Generate audio versions of news articles in real-time.
  3. IVR Systems: Power interactive voice response systems with fast, efficient TTS.
  4. Bulk Audio Generation: Create large volumes of audio content for various applications.
  5. E-learning Platforms: Generate audio content for online courses efficiently.

​
candy.two model

Our flagship model focused on high-quality, natural-sounding speech synthesis, making it ideal for applications requiring premium voice output. View all candy.two speakers

​
Use Cases

  1. Virtual Assistants: Create lifelike AI assistants with natural-sounding voices.
  2. Audiobook Production: Automate the creation of audiobooks with high-quality narration.
  3. Podcast Creation: Produce professional-sounding podcast episodes from written scripts.
  4. Gaming: Develop dynamic, responsive NPC dialogues in video games.
  5. Accessibility Tools: Improve web accessibility with high-quality text-to-speech for visually impaired users.

​
swift.call model

Our specialized model designed specifically for voicebot applications, featuring real-time voice cloning and streaming capabilities. View all swift.call speakers

​
Key Features

  1. Run-time Voice Cloning: Clone voices in real-time for dynamic voice adaptation
  2. Word-level Timestamps: Get precise timing information for each word in the generated speech
  3. Streaming Support: Real-time audio streaming for interactive voicebot applications
  4. Low Latency: Optimized for real-time voice interactions

​
Use Cases

  1. Voicebots: Create interactive voice assistants with real-time voice cloning
  2. Call Centers: Power automated customer service systems with natural-sounding voices
  3. Interactive Voice Response: Build dynamic IVR systems with streaming support
  4. Real-time Voice Applications: Develop applications requiring immediate voice feedback

​
Model Comparison

Featureneo.one candy.two swift.call
Processing SpeedUltra-fastStandardReal-time
Voice QualityGood qualityHighest fidelityHigh quality
Resource UsageOptimized (lower)StandardOptimized
Language Support10+ languages8+ languagesHindi & English
Emotion ControlBasicAdvancedBasic
Batch ProcessingOptimizedSupportedStreaming
ArchitectureNon-autoregressiveTransformer-basedVoicebot-optimized
Use CaseEfficient, large-scale TTSHigh-quality, premium TTSReal-time voicebot TTS
Voice CloningNoNoYes
Word TimestampsNoNoYes

​
Choosing the Right Model

  • Choose neo.one when:
    • Processing speed is a top priority
    • You’re generating large volumes of audio content
    • You need to optimize resource usage
    • Batch processing is required
  • Choose candy.two when:
    • You need the highest quality voice output
    • Advanced emotion and emphasis control is required
    • You’re working on applications requiring premium voice quality
    • Transformer-based architecture benefits are desired
  • Choose swift.call when:
    • You’re building voicebot applications
    • Real-time voice cloning is required
    • Streaming support is essential
    • Word-level timestamps are needed
    • You need Hindi or English voice support
For guidance on selecting the best model for your specific use case, please contact our support team at [email protected].