OpenAI TTS
audioby OpenAI · Updated February 15, 2026
OpenAI's Text-to-Speech (TTS) models convert text into natural-sounding speech using 6 distinct voice options (Alloy, Echo, Fable, Onyx, Nova, Shimmer). Available in standard and HD quality, the models support real-time streaming, 24+ languages, and produce clear, expressive speech suitable for narration, accessibility, and conversational AI applications.
Best For
Prompting Tips
- 1Choose the right voice: Alloy (neutral), Echo (warm), Fable (storytelling), Onyx (deep), Nova (bright), Shimmer (gentle)
- 2Use HD model for highest quality, standard for lower latency
- 3Punctuation controls pacing — use commas, periods, and ellipses deliberately
- 4Supports 24+ languages automatically based on input text
Syntax & Constraints
Text input via API. 6 voice options. HD and standard quality. Real-time streaming supported. Max 4096 characters per request.
Build Prompts for OpenAI TTS
Other ChatGPT Models
OpenAI's fastest multimodal flagship model.
Small, fast, and affordable model for lightweight tasks.
High-capability model with vision and 128K context.
OpenAI's image generation model with strong text understanding.
OpenAI's native image generation built into ChatGPT.
OpenAI's text-to-video model generating realistic scenes.
OpenAI's evolved video model with character consistency.