Developer Settings

The Developer Settings section allows advanced users to customize the providers used for core AI services such as Speech-to-Text (STT), Text-to-Speech (TTS), and Language Models (LLM/GPT). By adjusting these, you can fine-tune performance, cost, latency, and language capabilities according to your project’s needs.

🔹 What You Can Configure

1. STT (Speech-to-Text Provider)
   • Converts spoken words from callers into text.
   • Options may include Google Cloud Speech, Deepgram, AssemblyAI, etc.
   • Factors to consider:
     • Accuracy (especially for medical/legal use cases)
     • Supported languages & accents
     • Latency & pricing per minute
2. TTS (Text-to-Speech Provider)
   • Converts the AI’s text response back into spoken voice.
   • Options may include ElevenLabs, Azure TTS, Google Cloud TTS, etc.
   • Factors to consider:
     • Voice naturalness
     • Available accents, gender, tone
     • Multilingual support
     • Per-character or per-second pricing
3. LLM / GPT (Large Language Model Provider)
   • Powers the “thinking” of your AI agent.
   • Options may include OpenAI GPT-4o / GPT-4o-mini, Anthropic Claude, or other integrated models.
   • Factors to consider:
     • Reasoning ability (complex vs. simple conversations)
     • Latency (response speed for live calls)
     • Cost per token
     • Context length (important for long conversations)

🔹 Why Customize Providers?

• Cost Control → Balance between high-end natural voices (more expensive) vs. budget options.
• Performance Optimization → Reduce latency by choosing faster STT/LLM combos.
• Language Coverage → Ensure the provider supports your target audience’s language and dialect.
• Voice Branding → Select TTS voices that match your brand identity.
• Flexibility → Swap providers without rebuilding your entire agent.

🔹 Example Configurations