- Supported Audio Codecs
- MP3, WAV, Opus, FLAC, ulaw (8kHz), alaw
- Audio Sample Rates
- 8kHz, 16kHz, 24kHz, 48kHz
- Minimum Audio Bit Depth
- 16-bit (8-bit optional for legacy systems)
- Noise Cancellation
- Yes
- Echo Cancellation
- Yes
- Concurrent Call Capacity
- Platform-dependent; typically 100-10,000+ simultaneous calls
- SLA Uptime Guarantee
- 99.5%-99.99% (vendor-specific)
- Geographic Deployment Options
- Multi-region cloud, on-premises, hybrid
- LLM Context Window
- 4K-200K tokens (varies by platform)
- Supported Languages
- Typically 50-100+ languages with regional variants
- Code-switching Support
- Platform-dependent; most support multi-language conversations
- Average Response Latency (P95)
- <3.5 seconds
- STT Latency Component
- 200-600ms depending on speech length
- LLM Inference Latency
- 500-1500ms depending on response length
- TTS Synthesis Latency
- 100-500ms for typical responses