Amazon Polly launches Bidirectional Streaming API to cut text-to-speech latency for conversational AI apps

Amazon Web Services launched a Bidirectional Streaming API for Amazon Polly, its text-to-speech service, allowing developers to send text and receive synthesized audio simultaneously over a single HTTP/2 connection rather than waiting for a complete text input before synthesis begins. In internal benchmarks using 970 words of prose, the new API processed audio 39% faster than the traditional approach and reduced API calls from 27 down to one, by streaming text word-by-word as an LLM generates it instead of waiting for full sentences. The API is now generally available and supports most major AWS SDKs including Java, JavaScript, .NET, Go, Ruby, Rust, and Swift, with Python and CLI support not yet included.