Amazon Polly
Amazon Web Services
Amazon Polly is the most cost-effective production-grade TTS API available, with Standard voices at $4/1M characters — roughly 3% of OpenAI Realtime's effective per-minute cost. Four pricing tiers: Standard ($4/1M, robotic quality), Neural ($16/1M, significantly better), Generative ($30/1M, English-only LLM-based, launched September 2025), and Long-form ($100/1M, optimized for audiobooks). Free tier: 5M characters/month for the first 12 months. 60+ languages and 100+ voices. Best for AWS-centric teams, high-volume IVR, and applications where cost at scale is the primary constraint and voice quality is secondary.
Free: 5M chars/month for first 12 months. Standard: $4/1M chars. Neural: $16/1M. Generative: $30/1M (English only). Long-form: $100/1M. Pay-as-you-go.
Related platforms
Cartesia
Cartesia
Ultra-low latency voice AI for real-time agents — Sonic 2 at sub-100ms, instant cloning from 3 seconds.
Azure AI Speech
Microsoft
Enterprise TTS with 500+ voices across 140+ languages — Custom Neural Voice and Fortune 500 compliance.
Deepgram
Deepgram
Enterprise speech-to-text and voice AI platform.