Unreal Speech: Text-to-Speech API

What is Unreal Speech?

Unreal Speech is a text-to-speech API built for developers and businesses that need to convert written content into natural-sounding audio at scale. It handles everything from short real-time snippets to long-form audio files containing hundreds of thousands of characters. Teams use this text-to-speech API to add voice output to apps, automate audio content production, and reduce the cost of speech synthesis compared to larger providers.

Features & Benefits

• Text-to-audio synthesis – Convert text to spoken MP3 audio across three input sizes: up to 1,000 characters via synchronous streaming, up to 3,000 characters with timestamp output, and up to 500,000 characters via asynchronous task processing.
• Real-time audio streaming – Deliver synthesized speech in as little as 0.3 seconds using the /stream endpoint, suitable for live applications and low-latency voice interfaces.
• Long-form audio generation – Process up to 500,000 characters per task asynchronously, with approximately 10 hours of audio producible in around 15 minutes.
• Per-word and per-sentence timestamps – Return precise timing data for each word or sentence alongside generated audio, enabling synchronized text highlighting and caption display.
• WebSocket streaming with timestamps – Stream audio and word-level timing data simultaneously over a WebSocket connection for real-time word-by-word highlighting use cases.
• Voice library – Access 48 voices across 8 languages including US English, UK English, Mandarin Chinese, Hindi, Spanish, Portuguese, Japanese, French, and Italian.
• Audio output controls – Adjust speech speed (−1.0 to 1.0), pitch (0.5 to 1.5), bitrate (up to 320k), and codec (MP3 or PCM) per request.
• Callback URL support – Receive a server ping when an asynchronous synthesis task completes, avoiding the need to poll for status.
• Multi-language SDK support – Integrate the text-to-speech API using Python, Node.js, React Native, or cURL with provided code samples and SDKs.
• Commercial usage rights – Use generated audio in commercial products; paid plans require no attribution.

Real-World Applications

Developers building read-aloud features for e-learning platforms or reading apps can use the text-to-speech API to convert lesson text or articles into audio on demand. The low-latency /stream endpoint makes it practical for real-time playback triggered by user actions, such as tapping a “listen” button on a mobile app. Timestamp data allows the app to highlight each word as it plays, which can support learners who benefit from visual reinforcement.

Content teams producing high volumes of audio for podcasts, audiobooks, or narrated video scripts may find the asynchronous /synthesisTasks endpoint useful. It can process book-length text in a single request and return a downloadable MP3, reducing the manual work involved in studio recording or file management. Businesses converting large document libraries into audio format can automate that pipeline entirely through the API.

Interactive voice applications — such as customer service bots, navigation systems, or notification readers — benefit from the sub-second response time the streaming endpoint provides. Developers can pass dynamic text at runtime and receive speech output fast enough to maintain a natural conversational flow. The adjustable pitch and speed parameters let teams tune voice output to match a product’s tone without sourcing a separate voice actor.

Publishers and media companies looking to add audio versions of their articles might use the text-to-speech API to generate multilingual audio from the same source text. With voices available in Spanish, French, Hindi, Japanese, and other languages, a single integration can serve readers across different regions without managing separate localization workflows.