TTS Provider Interface

Unify streaming TTS workflows by using a provider-agnostic interface that standardizes audio synthesis across Deepgram, AWS Polly, and Google Cloud. It removes the need to implement provider-specific adapters separately while adding latency-aware streaming.

Switch between Deepgram, AWS Polly, and Google Cloud TTS with minimal changes by using a provider-agnostic interface. This interface standardizes audio synthesis and latency tracking, eliminating the need for provider-specific adapters.

Track first-byte latency for real-time audio synthesis by initiating streaming with the tts.synthesize tool. It provides real-time audio chunks alongside latency metrics, enabling low-latency synchronous audio output.

The TTS interface supports Twilio-compatible playback by enforcing mulaw 8kHz mono output formatting. This ensures the streaming audio synthesis output is correctly formatted for telephony applications.

Benchmark and cancel streaming audio synthesis using the tts.benchmark and tts.cancel tools. The interface also provides tts.status for monitoring, alongside tts.synthesize for initiating the real-time audio generation.

Configure voice selection for cross-provider text to speech synthesis by providing a text input and voice config to the tts.synthesize tool. This initiates streaming audio output formatted as mulaw 8kHz mono.