Unreal Speech

Unreal Speech is a developer-first text-to-speech API built around one idea: natural-sounding speech at a fraction of the usual per-character cost.

Overview

Unreal Speech focuses on doing one thing well — fast, inexpensive speech synthesis through a clean API. Its streaming endpoint returns audio in about 300 milliseconds for short passages, which makes it suitable for real-time applications, while asynchronous endpoints handle long-form jobs of up to 500,000 characters for audiobooks and bulk narration. A WebSocket endpoint streams word-level timestamps, so developers can build karaoke-style highlighting and synchronized captions. The catalog spans dozens of voices across major languages including English, Spanish, French, Hindi, Mandarin, and Japanese, with controls for speed, pitch, and bitrate.

Key Features

• Streaming API with ~0.3s response for passages up to 1,000 characters
• Word-level timestamps over WebSocket for highlighting and captions
• Async synthesis tasks for long documents up to 500,000 characters
• Voice, speed, pitch, and bitrate controls with MP3 and PCM output
• Generous free tier and volume pricing that drops as usage grows
• Unused characters on paid plans roll over to the next month
• Commercial use allowed on all plans (attribution required on free)

💳 Pricing

Free plan with 250,000 characters per month (about 6 hours of audio). Basic is $49/month for 3M characters, Plus is $499/month for 42M, Pro is $1,499/month for 150M, and Enterprise is $4,999/month for 625M characters, with overage rates falling from $16 to $8 per million characters.

Who Is It For

Developers and startups that need TTS at serious volume — listening apps, audiobook pipelines, e-learning platforms — and want transparent, predictable per-character pricing.