What is a Text-to-Speech API and How Does It Work?

By WordCaster editorial team · Updated 2026-11-18

This page explains how cloud text-to-speech APIs convert written content into audio files, focusing on the WordCaster platform built for indie developers and no-code builders. You will learn how simple JSON requests replace complex infrastructure, enabling you to generate speech without managing servers or licensing deals. We cover the technical workflow from input processing to audio delivery, ensuring you understand the exact steps needed to integrate TTS into your projects.

We break down the core components, including the 8 voice presets, language support, and audio formats like MP3 and WAV. The guide details the API structure, pricing model, and integration steps, helping you move from sign-up to your first audio file efficiently. Whether you are building a mobile app or automating voiceovers, this overview provides the specific facts you need to implement TTS with minimal friction and transparent costs.

What is a Text-to-Speech API and How Does It Work?

Frequently asked questions

What is a TTS API?

A cloud text-to-speech API lets you send text and receive an audio file via a simple POST request. It uses an OpenAI-compatible endpoint, allowing you to use standard SDKs by changing the base URL. This provides a straightforward way to convert written content into spoken audio without managing complex infrastructure.

How is audio generated?

You send a JSON body with your text, voice selection, and format preferences to the speech endpoint. The system processes the request and returns the audio file. For longer texts, you split the input into 4096-character chunks, generate each part, and join the resulting audio files.

What is the difference between tts-1 and tts-1-hd?

Both models use the same engine and incur the same price of $10 per million characters. The distinction lies in the audio quality profile, but functionally they operate identically within the API. You can choose either based on your specific quality needs without any change in cost or structure.

Can the API generate audio in real-time?

No, the API does not support streaming or WebSocket connections for real-time partial audio. Requests are processed sequentially, so under load, a request may wait in a queue. Typical generation speed is roughly real-time for long texts, but short phrases may take a few seconds to complete.

Does the API support SSML?

No, the API does not support SSML (Speech Synthesis Markup Language). It accepts plain text input up to 4096 characters. You cannot use SSML tags to control pacing, pronunciation, or emphasis directly within the request body, relying instead on the natural intonation of the selected voice presets.

What languages are supported?

The API supports 23 languages, including English, Spanish, French, German, Japanese, and Arabic. Any voice can speak any of these languages. If the language parameter is omitted, the system guesses the language, defaulting to English for Latin script text. Vietnamese, Thai, and Ukrainian are not supported.

How many characters can I send?

Each request can contain up to 4096 characters of text. If your input exceeds this limit, you must split it into smaller parts. The API does not automatically handle long texts; you need to manage the splitting logic and join the resulting audio files yourself.

Is the audio stored after generation?

No, the generated audio is returned in the response and is not stored by the service. Your input text is used only to produce the audio and is not used for training. Only request metadata, such as time, character count, and status, is kept for billing purposes.

Can I use the API for commercial projects?

Yes, the API is designed for indie developers and no-code builders to integrate into commercial projects. There are no restrictions on usage type. You pay only for the characters processed, with no monthly plans or minimum spend requirements, making it suitable for both small and large-scale applications.

What happens if the API fails?

If synthesis fails, the charge is returned in full. You are not billed for failed requests. This ensures that you only pay for successfully generated audio files. You can retry the request or adjust parameters like speed or voice without incurring additional costs for the failed attempt.

How do I integrate the API into my app?

Send a POST request to the speech endpoint with your API key in the Authorization header. Use the standard OpenAI SDK structure by setting the base URL to the API domain. Include your text, voice, and format in the JSON body. The response will contain the audio file.

Do I need a developer to use it?

While a developer is helpful for integration, the API is designed for no-code builders too. You can use the provided curl examples and simple JSON structures. Many no-code platforms can handle HTTP requests, allowing you to connect the API to your workflows without writing extensive code.

Is the API free to use?

No, the API is pay-as-you-go at $10 per million characters. New users signing in with Google receive a $0.50 free trial valid for 7 days. Email-based accounts start without trial credit. You can top up your balance using USDT on the TRON network (TRC-20).

How accurate is the language detection?

The API guesses the language from the script if not specified, defaulting to English for Latin text. For non-English Latin-script text, it is recommended to explicitly set the language parameter. Detection relies on character scripts, so accurate language specification ensures the correct voice pronunciation and intonation.

Definition

Sound wave illustration for: Definition

A Text-to-Speech (TTS) API converts written text into audible speech files via a cloud endpoint. Instead of installing software, developers send a JSON payload containing text and voice settings, receiving an audio file (MP3, WAV, etc.) in return. This enables seamless audio integration into applications without managing complex infrastructure.

What is TTS

TTS synthesizes human-like speech from digital text. It is a core component for adding voice to apps, websites, or automated systems. The process transforms characters into sound waves using trained neural models, ensuring consistent quality across different devices and platforms.

Cloud-based generation

All processing happens on remote servers. You send text via an HTTP POST request to an endpoint like https://api.playtoto155.com/v1/audio/speech and receive the audio file. This eliminates the need for local hardware acceleration or large model downloads, keeping your application lightweight and scalable.

API vs local engines

Unlike local engines that require installation and consume device resources, a cloud API offers instant updates and consistent voice quality. You pay only for characters used, avoiding upfront software costs. This pay-as-you-go model suits indie developers and no-code builders who need predictable expenses without managing server maintenance.

How it works

Sound wave illustration for: How it works

A Text-to-Speech API converts written text into spoken audio through a cloud service. Instead of running on your device, the processing happens on remote servers, allowing you to send a simple request and receive an audio file without managing complex infrastructure.

The system processes your input text, synthesizes it using neural models, and delivers the result as a downloadable audio file. This workflow is designed for developers who need reliable audio generation without building voice engines from scratch.

Text input processing

You send plain text to the API endpoint. The system accepts up to 4096 characters per request, including spaces and punctuation. If your text exceeds this limit, you must split it into smaller parts and join the resulting audio files later.

The service guesses the language based on the script if you don't specify one, though setting the language explicitly is recommended for non-English Latin text. It supports 23 languages, including Arabic and Turkish, ensuring accurate pronunciation for diverse content.

  • Max 4096 characters per request
  • Spaces and punctuation count as characters
  • Automatic language detection or manual ISO code

Neural voice synthesis

The API uses neural models to generate human-like speech. You choose from eight voice presets, such as the warm narrator 'Aria' or the deep documentary voice 'Orion'. These voices are based on licensed recordings from LibriVox, ensuring high quality without celebrity voice imitation.

You can adjust the speed between 0.25 and 4.0 to fit your needs. The generation takes roughly real-time for long texts, with short phrases taking a few seconds. Requests are processed sequentially, so high traffic may cause slight delays.

  • 8 voice presets available
  • Speed adjustable from 0.25x to 4.0x
  • Based on licensed LibriVox recordings

Audio file delivery

Once synthesis is complete, the API returns the audio file directly in the response. You can choose from MP3, WAV, Opus, FLAC, or PCM formats depending on your quality and size requirements. The file is ready for immediate use in your application.

Each generated file contains an inaudible watermark to mark it as synthetic speech. The input text is not stored or used for training; only request metadata like time and character count is kept for billing purposes.

  • Formats: MP3, WAV, Opus, FLAC, PCM
  • Inaudible synthetic watermark included
  • Audio returned in response, not stored

Technical components

Sound wave illustration for: Technical components

The API exposes a single REST endpoint for synthesis. Developers send standard HTTP POST requests to generate audio files, ensuring compatibility with existing tools and libraries without requiring custom infrastructure.

Requests use a JSON body containing parameters like model, input text, voice selection, and audio format. This structure mirrors the OpenAI speech endpoint, allowing developers to swap base URLs in their SDKs with minimal code changes.

The response delivers the audio file directly in the requested format, enabling immediate playback or storage in cloud buckets.

Endpoint structure

All synthesis requests target POST https://api.playtoto155.com/v1/audio/speech. This single endpoint handles all audio generation tasks, simplifying integration for no-code builders and indie developers who need a reliable, predictable interface.

Authentication is handled via a Bearer token in the Authorization header. Each account holds one active key, ensuring secure and traceable usage for billing purposes.

  • Base URL: https://api.playtoto155.com/v1
  • Method: POST
  • Content-Type: application/json

JSON request body

The request body accepts model (tts-1 or tts-1-hd), input (up to 4096 characters), voice, response_format, and speed. Optional language codes can override automatic detection for Latin-script texts.

Text exceeding 4096 characters must be split into multiple requests. The client application is responsible for joining the resulting audio files.

  • model: tts-1 or tts-1-hd
  • input: string, max 4096 chars
  • voice: 8 preset options
  • response_format: mp3, wav, opus, flac, pcm

Response handling

The API returns the generated audio file directly in the response body. If the request fails, no charge is incurred, and the error details are provided in the response.

Clients should handle the binary audio data or stream it directly to a player. The speed parameter allows adjustment between 0.25 and 4.0, affecting both generation time and final playback duration.

  • Returns binary audio data
  • Failed requests are free
  • Supports 0.25–4.0x speed

Voices and models

Sound wave illustration for: Voices and models

WordCaster provides eight distinct voice presets, ranging from warm narrators to energetic promo voices. Each voice is derived from licensed recordings, ensuring high-quality synthetic speech without the need for complex infrastructure or custom voice cloning.

The engine offers two model variants: 'tts-1' and 'tts-1-hd'. Both utilize the same underlying engine and pricing structure, allowing you to choose between standard efficiency and enhanced clarity based on your project's specific audio requirements.

Voice presets

The API includes eight voice options, including female voices like Aria and Nova, and male voices such as Felix and Atlas. These presets cover various tones, from calm instructional styles to deep documentary narration.

All voices can speak any of the supported 23 languages, ensuring consistent quality across diverse linguistic contexts without requiring separate voice models for each region.

  • Aria: warm, clear narrator
  • Nova: bright, upbeat
  • Felix: energetic, promo
  • Atlas: neutral, news

HD vs standard models

Both 'tts-1' and 'tts-1-hd' models use the same engine and incur identical charges of $0.01 per 1,000 characters. The 'hd' variant offers enhanced audio fidelity, while the standard model provides efficient processing for bulk generation.

Since pricing and base capabilities are identical, developers can switch between models without adjusting their cost expectations or request structures.

Voice characteristics

Generated audio files contain an inaudible watermark, clearly marking them as synthetic speech. This ensures transparency without affecting the listener's experience.

Each voice maintains consistent pacing, derived from specific licensed recordings, providing a reliable audio brand identity for your applications.

Language support

Sound wave illustration for: Language support

WordCaster supports 23 languages, including English, Spanish, French, German, Japanese, and Arabic. Every voice preset can speak any of these languages, allowing you to build multilingual applications without managing separate voice models or regional configurations.

You do not need to select a specific voice for a language; the system handles the localization through the voice model itself, ensuring consistent quality across all supported tongues.

Multilingual capabilities

The API supports 23 languages, ranging from common ones like English and Spanish to others like Swahili and Turkish. Any of the eight voice presets can produce any of these languages, giving you flexible voice options regardless of the target language.

Language detection

If you do not specify a language, the API guesses it from the text script. For example, Latin script text defaults to English. However, for non-English Latin-script text, you should explicitly set the language parameter to avoid incorrect pronunciation.

  • Supported: Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish.

Script-based guessing

The system uses the character script to infer the language when the language field is omitted. This works well for standard cases but may fail if the text contains mixed scripts or ambiguous characters, so explicit language codes are recommended for precision.

Audio formats

Sound wave illustration for: Audio formats

The API accepts requests for audio in five distinct formats, allowing you to choose the right balance between file size and fidelity for your specific workflow. Each format is delivered as a downloadable binary file within the standard JSON response.

MP3 compatibility

MP3 is the default and most widely supported format, making it ideal for web players, mobile apps, and general content narration. It offers efficient compression that keeps file sizes manageable without significant loss in perceived quality for speech.

  • Best for web embedding and mobile playback
  • Common lossy format

WAV for quality

WAV provides uncompressed, high-quality audio suitable for professional editing or archival purposes. It preserves the full fidelity of the generated voice, ensuring no data is lost during transmission or storage.

  • Uncompressed lossless audio
  • Ideal for post-production editing

Opus and FLAC options

Opus offers superior compression efficiency compared to MP3, maintaining high quality at lower bitrates. FLAC provides lossless compression, reducing file size while retaining all original audio data, making it perfect for storage-efficient backups.

  • Opus: High efficiency, low bitrate
  • FLAC: Lossless, space-efficient storage

Integration steps

Sound wave illustration for: Integration steps

Integration begins by securing your access credentials. Visit the /get-api-key/ page to register via Google or email, where you will receive a unique API key. This key authenticates your requests and tracks your usage against your available balance.

Once configured, you are ready to convert text into audio using standard HTTP methods or the OpenAI SDK with a adjusted base URL.

API key generation

Register on the /get-api-key/ page using Google or email. The dashboard displays your active key, current balance, and a ready-to-use cURL example. Keys are unique per account and can be rotated immediately if needed.

Sending the request

Construct a POST request to https://api.playtoto155.com/v1/audio/speech. Include your key in the Authorization header and provide the text, voice, and format in the JSON body. The API processes requests sequentially, so longer texts may queue briefly.

Receiving the audio

Upon success, the endpoint returns the audio file directly in the response body. Failed requests incur no charge. For texts exceeding 4096 characters, split the input, process each part, and join the resulting audio files.

Integration steps

Use cases

Sound wave illustration for: Use cases

Text-to-speech APIs enable developers to embed audio directly into applications without managing complex infrastructure. By converting text into audio files via simple HTTP requests, indie developers can add voice capabilities to their projects with minimal code.

This approach is particularly useful for no-code builders who need to generate voice content dynamically. The pay-as-you-go model ensures costs scale with usage, making it ideal for projects with variable traffic patterns.

Mobile app audio

Developers can integrate TTS into mobile apps to read notifications, provide accessibility features, or narrate content. The API supports multiple languages and voices, allowing for localized audio experiences without downloading large voice packs.

Users can generate audio on demand, ensuring the content is always up-to-date. The lightweight nature of the API responses makes it suitable for devices with limited processing power.

  • Generate audio for push notifications or in-app messages
  • Provide accessibility options for visually impaired users
  • Localize content for different regions with various languages

Web content narration

Websites can use TTS to offer audio versions of articles, blog posts, or documentation. This enhances user engagement and allows users to consume content hands-free.

Since the API returns standard audio formats like MP3 and WAV, integrating audio players into web pages is straightforward. The JSON-based request structure makes it easy to trigger audio generation from user interactions.

  • Convert blog posts or articles into audio files
  • Enhance accessibility for content consumers
  • Reduce bounce rates by offering alternative content formats

Automated voiceovers

Content creators can automate the production of voiceovers for videos, podcasts, or presentations. By sending text through the API, they can generate high-quality audio files quickly.

The variety of voice presets allows for diverse tonal options, from warm narrators to energetic promo voices. This flexibility ensures that the audio matches the intended mood of the content.

  • Generate voiceovers for video content
  • Create audio for e-learning modules
  • Produce consistent voice branding across multiple projects

Limitations

Sound wave illustration for: Limitations

Text-to-Speech APIs convert written content into audio files via a simple request-response cycle. While convenient, these systems have specific boundaries regarding input length, formatting capabilities, and delivery methods that developers must account for in their architectures.

Character limits

Each API request accepts up to 4,096 characters. For longer texts, you must split the input into chunks, generate audio for each part, and join the files. Short phrases may take a few seconds to process, while longer texts generate at roughly real-time speed.

No SSML support

The API does not support Speech Synthesis Markup Language (SSML). You cannot control pronunciation, pauses, or emphasis through tags. Voice selection is limited to eight presets with fixed characteristics, such as 'aria' for narration or 'atlas' for news, without custom emotion or timing adjustments.

No streaming

Audio is delivered as a complete file in the response body, not as a real-time stream. You cannot receive partial audio chunks while generation is ongoing. This means the entire text must be processed before the audio data is available, which affects user experience in live applications.

Data handling

Sound wave illustration for: Data handling

Your input text is used exclusively to generate the audio file and is never used for training the model. The system processes your characters to synthesize speech without retaining the source content for future voice improvement or analysis.

Generated audio is returned directly in the API response and is not stored on our servers after delivery. This ensures that your content remains transient, reducing long-term data retention risks for indie developers managing sensitive or proprietary text.

We keep request metadata—such as timestamp, character count, and status—for billing accuracy. This data helps verify usage against your balance but does not include the actual text or audio files themselves.

Input text usage

Input text is used solely for synthesis and is not retained for training purposes. The engine processes your characters to create audio without storing the original content for future model improvement.

  • Text is not used for training
  • No long-term content retention

Audio storage

Generated audio files are returned in the response and are not stored on our servers. Once the download completes, the audio data is effectively ephemeral.

This approach minimizes storage costs and ensures your audio files remain under your control without being archived in our database.

  • Audio is not stored
  • Returned in response only

Billing metadata

We retain request metadata, including time, character count, and status, strictly for billing verification. This data allows us to accurately charge your balance and troubleshoot issues without keeping the audio or text content.

Request metadata is kept for billing.

  • Time and status kept
  • Character count for billing

Getting started

Sound wave illustration for: Getting started

Getting started with WordCaster is straightforward for indie developers and no-code builders. The process begins with account creation, followed by a simple API key generation and a test request to verify integration.

Sign up process

Visit the /get-api-key/ page to create an account. You can sign in using Google for an instant $0.50 trial credit (valid for 7 days) or use an email and password. Accounts are specific to this site.

  • Google sign-up includes $0.50 trial credit
  • Email/password accounts start with zero balance
  • One active API key per account

First request

Use the provided API key to send a POST request to https://api.playtoto155.com/v1/audio/speech. The endpoint accepts JSON bodies similar to the OpenAI format, allowing you to specify the model, voice, and input text immediately.

  • Endpoint: POST /v1/audio/speech
  • Auth: Bearer tts-... key
  • JSON body with text and voice

Testing integration

After sending the request, the API returns an audio file. You can verify the output by playing the returned file or checking the response status. The dashboard on /get-api-key/ shows your usage and balance in real time.

  • Check dashboard for usage stats
  • Verify audio file download
  • Monitor balance after request

Comparison to alternatives

Sound wave illustration for: Comparison to alternatives

WordCaster offers a familiar interface for developers already using OpenAI’s ecosystem, but operates as an independent service with its own pricing structure and voice engine. It is not affiliated with OpenAI, despite supporting the same endpoint shape.

OpenAI compatibility

The API uses the same request schema as OpenAI’s tts-1 endpoints, allowing you to swap the base_url to https://api.playtoto155.com/v1 and reuse existing client code. This compatibility simplifies integration for indie developers who already have OpenAI SDKs configured.

  • Identical JSON request body structure
  • Supports tts-1 and tts-1-hd models
  • Works with official OpenAI SDKs

ElevenLabs differences

Unlike ElevenLabs, WordCaster does not offer custom voice cloning or a dedicated mobile SDK. It provides eight fixed voice presets derived from licensed LibriTTS-R recordings, focusing on stable, consistent narration rather than dynamic emotional variation or community-driven voice libraries.

  • No voice cloning or custom uploads
  • Fixed set of 8 licensed voices
  • No mobile SDK available

Google Cloud differences

Google Cloud TTS requires complex infrastructure setup and enterprise contracts, whereas WordCaster uses a simple pay-as-you-go model with no minimum spend or monthly subscriptions. It supports fewer languages (23 vs. Google’s extensive list) but offers a more straightforward, developer-first API experience.

  • No enterprise contracts or SLAs
  • Pay-as-you-go billing only
  • Supports 23 languages

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. Each voice is based on licensed recordings from LibriVox, ensuring clear, professional audio without licensing complexity.

Avatar of the voice aria
ariafemale
Avatar of the voice nova
novafemale
Avatar of the voice luna
lunafemale
Avatar of the voice sage
sagefemale
Avatar of the voice felix
felixmale
Avatar of the voice milo
milomale
Avatar of the voice atlas
atlasmale
Avatar of the voice orion
orionmale

23 languages

Any voice speaks any of these languages.

Sources and further reading

  1. Wikipedia
  2. ISO 639 language codes
  3. curl documentation
  4. IETF RFC 7845: Ogg Opus
  5. W3C WAI: making audio and video accessible
  6. LibriTTS-R corpus (OpenSLR 141)
  7. openai-python on GitHub
  8. openai-node on GitHub
  9. Node.js documentation
  10. OpenAI API reference: create speech
  11. FFmpeg documentation
  12. Python documentation

More on this site