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

By Readly API editorial team · Updated 2026-11-18

This page explains how a Text-to-Speech (TTS) API works, tailored for indie developers integrating audio into SaaS notifications and interfaces. We break down the technical workflow, from sending text to receiving audio files, without enterprise overhead or hype. You will learn how our simple REST API fits into your existing stack using familiar tools.

Read on to understand voice selection, latency, pricing, and data handling. We cover integration methods, limitations like the lack of streaming, and how to get started with a free trial. Whether you need accessible content or automated alerts, this guide provides the concrete facts you need to build reliable audio features.

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

Frequently asked questions

What is a TTS API?

A Text-to-Speech API is a cloud service that converts written text into audio files. You send a POST request with your text, and the API returns an audio file like MP3 or WAV. It is designed for developers who need to add voice capabilities to applications without managing complex infrastructure.

How does text to speech work?

You send text to our endpoint at POST https://api.playpx9.com/v1/audio/speech. The system processes the input using licensed voice models and returns the generated audio file. You can specify the voice, language, and format. The service handles the synthesis server-side, so you just receive the final audio result.

Is a TTS API hard to integrate?

Integration is straightforward because we use the same request shape as the OpenAI speech endpoint. You can use the official OpenAI SDKs by simply changing the base_url to our API domain. We also provide ready curl examples on your dashboard, making it easy to start coding immediately.

Do I need special hardware?

No special hardware is required. Since the synthesis happens on our cloud servers, your device only needs to handle HTTP requests and receive the resulting audio file. This works on any machine capable of running standard web applications, regardless of CPU or GPU power.

Can I use TTS for commercial projects?

Yes, you can use the generated audio in commercial projects. The voices are based on licensed recordings from LibriTTS-R and LibriVox. Each file includes an inaudible watermark to mark it as synthetic speech. There are no additional licensing fees, though each file carries an inaudible watermark marking it as synthetic speech.

Is the audio quality high?

We offer two models: tts-1 and tts-1-hd, which use the same engine and price. The voices are based on licensed recordings, ensuring clear and natural sound. You can choose from eight distinct voice presets, including options for narration, news, and conversational styles, to fit your specific project needs.

Does it support multiple languages?

We support 23 languages, including English, Spanish, German, French, Japanese, and Chinese. Any voice can speak any of these languages. If you do not specify a language, the system guesses it from the script, but setting it explicitly ensures accuracy, especially for non-English Latin-script text.

Can I control the speed?

Yes, you can adjust the speech speed using the speed parameter in your request. The acceptable range is from 0.25 to 4.0. This allows you to slow down the narration for clarity or speed it up for concise updates, giving you full control over the playback rate.

Is there a character limit?

Each request supports up to 4096 characters of input text. For longer texts, you should split them into parts of this size and join the resulting audio files. This limit ensures efficient processing while maintaining high quality for each segment of your content.

Do I need to download the voice?

No, you do not need to download or install any voice packs. The voices are hosted on our cloud infrastructure. You simply select a voice ID in your API request, and the system generates the audio using the appropriate model instantly without any local storage requirements.

Can I clone a voice?

No, we do not offer custom voice cloning. You must choose from our eight preset voices: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, or Orion.

Is SSML supported?

No, our API does not support SSML (Speech Synthesis Markup Language). You send plain text, and the system handles the pronunciation and pacing automatically. If you need precise control over pauses or emphasis, you may need to split your text into multiple requests with specific timing.

How much does it cost?

Pricing is $10 per 1 million characters, or $0.01 per 1,000 characters. Spaces and punctuation count as characters. Failed requests are free, with charges returned in full. You can top up your balance using USDT on the TRON network, with no monthly plans or minimum spend.

Is there a free tier?

Yes, signing in with Google gives you a free trial of $0.50 credit, valid for 7 days. This covers about 50,000 characters. Accounts created with email and password do not receive this trial credit.

Definition of TTS API

Sound wave illustration for: Definition of TTS API

A Text-to-Speech API is a cloud service that converts written text into spoken audio. Instead of managing complex speech synthesis engines on your own servers, you send text to a remote endpoint and receive an audio file back. This approach keeps your infrastructure lightweight while delivering high-quality narration for your applications.

Cloud service that converts text to audio

Our API handles the heavy lifting of neural voice synthesis on our servers. You simply provide the text, and we return the audio. This model allows indie developers to integrate professional-grade voice output into SaaS notifications and interfaces without worrying about server maintenance or scalability during traffic spikes.

  • Scalable cloud infrastructure
  • No server maintenance needed
  • Consistent audio quality

Send text, receive audio file

The interaction is straightforward: you POST a JSON request containing your text to the endpoint. The system processes the input and returns the synthesized audio file in your chosen format. This simple request-response cycle makes it easy to embed voice features directly into your existing workflows or notification systems.

  • JSON-based requests
  • Instant audio delivery
  • Simple integration

No local software required

You do not need to install any specific software or voice packs on your end. Everything runs in the cloud, accessible via standard HTTP requests. This means your application can generate speech on any device or environment that can make a web request, from backend servers to mobile apps.

  • Zero local dependencies
  • Works on any platform
  • Easy to deploy

How Synthesis Works

Sound wave illustration for: How Synthesis Works

The API converts your input text into digital audio files using neural network models. You send a JSON payload with your desired voice and language settings, and the service processes the request to generate the corresponding sound wave.

Neural networks generate audio waveforms

Our engine synthesizes speech by predicting audio waveforms directly from text characters. This approach allows for natural-sounding output that adapts to the linguistic nuances of the input, whether it is a short notification or a longer paragraph of text.

Voices are based on licensed recordings

The eight available voice presets are derived from licensed recordings. Specifically, they utilize data from LibriTTS-R (CC BY 4.0), which includes readings from LibriVox public-domain audiobooks. This foundation ensures high-quality, human-like intonation without the need for custom voice training.

  • LibriTTS-R source data (CC BY 4.0)
  • Based on LibriVox public-domain readings
  • Eight distinct voice presets available

Watermark marks synthetic speech

Every audio file generated by the API contains an inaudible digital watermark. This subtle marker identifies the file as synthetic speech, which can be useful for transparency and content verification in your applications without affecting the listening experience.

Integration Methods

Sound wave illustration for: Integration Methods

REST API with JSON requests

Our API follows a straightforward REST pattern. You send a JSON payload containing the text, voice selection, and format preferences to the /v1/audio/speech endpoint. The service returns the synthesized audio file directly in the response, making integration into any stack that handles HTTP requests trivial.

Compatible with OpenAI SDKs

Since the request shape matches the OpenAI speech endpoint, you can use official OpenAI SDKs by simply changing the base_url to https://api.playpx9.com/v1 and using your API key. This allows you to leverage familiar code structures and libraries without learning a new client library from scratch.

Simple HTTP POST requests

For those preferring raw HTTP, a simple POST request with the correct headers and JSON body is all you need. Authentication is handled via a Bearer token in the Authorization header. This minimal approach ensures low overhead and fast integration for indie developers building SaaS notifications or interfaces.

Use Cases in SaaS

Sound wave illustration for: Use Cases in SaaS

For SaaS products, adding voice transforms static text into engaging audio experiences without requiring heavy infrastructure. Our API integrates directly into your notification systems, allowing you to send spoken updates to users via their preferred channels.

Automated phone call notifications

Generate voice messages for appointment reminders or system alerts. The API handles text-to-speech conversion, returning an audio file you can attach to your dialer service. This works well for keeping users informed without manual effort.

  • Send reminders for scheduled events
  • Deliver urgent system status updates
  • Provide voice confirmations for transactions

Audio alerts for dashboard updates

Enhance your dashboard with audible cues for critical events. When a task completes or a threshold is breached, trigger a short audio clip. This keeps developers and users aware of status changes even when not actively viewing the screen.

  • Notify on task completion
  • Signal data ingestion milestones
  • Alert on metric thresholds

Accessible content for users

Improve accessibility by converting long-form text into spoken audio. Users can listen to articles, documentation, or updates hands-free. With support for 23 languages, you can serve a global audience without hiring multiple voice actors.

  • Convert blog posts to audio
  • Read documentation aloud
  • Support multilingual user bases

Voice Customization

Sound wave illustration for: Voice Customization

Readly API offers eight distinct voice presets, allowing you to match your SaaS interface’s tone without complex configuration. Each voice is derived from licensed recordings, providing a consistent and recognizable audio identity for your application.

Choose from 8 preset voices

Select from female voices like Aria, Nova, Luna, and Sage, or male options including Felix, Milo, Atlas, and Orion. These presets cover a range of styles, from warm narration to energetic promo, ensuring your notifications fit your brand’s voice.

Adjust speed and language

Fine-tune playback speed between 0.25 and 4.0 to control pacing. You can also specify a language using its ISO code; if omitted, the system guesses the language from the text script, defaulting to English for Latin characters.

No custom voice cloning available

We do not offer custom voice cloning or the ability to upload your own voice samples. All audio is generated using our eight licensed presets, ensuring consistency and predictable quality across all your integrations.

Audio Formats

Sound wave illustration for: Audio Formats

You can choose from five output formats depending on your specific needs. MP3 is the standard for web and mobile applications due to its balance of quality and small file size. WAV and FLAC provide lossless audio for scenarios where maximum fidelity is required, such as archival purposes. OPUS and PCM are ideal for efficiency, reducing bandwidth usage while maintaining clear speech playback in constrained environments.

MP3 for web and mobile

MP3 is the recommended format for most integrations. It offers a good balance between audio quality and file size, making it ideal for web players and mobile apps where bandwidth efficiency matters. This format ensures smooth playback across various devices without significant data usage.

WAV and FLAC for quality

Choose WAV or FLAC when you need uncompressed, high-fidelity audio. WAV is widely supported and simple, while FLAC offers lossless compression for smaller file sizes without quality loss. These formats are suitable for premium features or when audio clarity is critical for user experience.

OPUS and PCM for efficiency

OPUS provides excellent compression for low-bandwidth scenarios, while PCM offers raw, uncompressed audio data. PCM is useful if you need to process the audio further before playback. These formats minimize data transfer costs and are efficient for high-volume automated notifications.

Latency and Speed

Sound wave illustration for: Latency and Speed

Generating audio takes time, but our system is optimized for efficiency. Long texts are processed at roughly real-time speed, meaning a 10-second audio file takes about 10 seconds to generate, though requests may wait in a queue under load. This makes the API suitable for background tasks or notifications where users don't expect instant playback.

Real-time generation for long texts

For extended content, generation speed matches playback speed. You can send large texts, and the system processes them sequentially. This predictable timing helps you estimate wait times accurately without complex calculations.

Queue processing under load

Requests are processed one after another. If many users generate audio simultaneously, your request may wait in a queue. This ensures stability but means short phrases might take a few seconds longer during peak times.

Approx 25 chars per second

We estimate speed at about 25 characters of text per second of generated audio. Short phrases still take a few seconds to process due to overhead. Plan your UI accordingly, showing loading states for these brief delays.

Data Handling

Sound wave illustration for: Data Handling

Your input text is used solely to generate the audio file and is not stored for training purposes. Once the synthesis is complete, the audio is returned directly in the API response and is not kept on our servers. This ensures your content remains yours without lingering copies in our database.

We retain request metadata—such as timestamp, character count, and status—exclusively for accurate billing and usage tracking. This lightweight data allows us to charge you precisely for what you use without retaining the actual voice content itself.

Input text not stored

The raw text you send is processed in memory to create the audio waveform. It is not archived or used to train our neural models, ensuring your specific content remains private and transient.

Audio returned in response

Generated audio is delivered immediately in the API response stream or file object. We do not store these audio files after delivery, so you must save them to your own infrastructure if you wish to keep them.

Metadata kept for billing

We log the time, character count, and status of each request to ensure accurate billing. This metadata is essential for calculating your usage and remains the only persistent data tied to your account activity.

Unlike input text, this metadata is kept for billing transparency and dispute resolution, but it does not contain the actual voice content or your original source text.

  • Timestamp of request
  • Character count processed
  • Request status (success/failure)

Cost Structure

Sound wave illustration for: Cost Structure

We charge $0.01 per 1,000 characters, which equals $10 per million characters. Spaces and punctuation count toward this total. You only pay for what you use, making it easy to predict costs for SaaS notifications or interface audio.

Pay per character used

Each request processes up to 4,096 characters. Longer texts are split and joined automatically. This predictable rate helps indie developers budget accurately without hidden fees or volume discounts.

  • 1 million characters = $10
  • 1,000 characters = $0.01
  • Spaces and punctuation count as characters

Failed requests are free

If synthesis fails, the charge is returned in full. You never pay for broken audio or timeouts. This ensures you only spend credit on successful deliveries.

No monthly subscription fees

There are no seats, minimums, or recurring plans. You top up with USDT (TRC-20) when needed. Balance does not expire, so you can pause usage without losing credit.

Cost Structure

Language Support

Sound wave illustration for: Language Support

Our API supports 23 languages, including Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. This broad coverage allows indie developers to localize SaaS notifications and interfaces for global audiences without managing multiple language-specific models.

The system is designed for flexibility: any of the eight preset voices can speak any of these supported languages. This means you can maintain a consistent brand voice across different regions, choosing the perfect tone—whether warm, upbeat, or calm—regardless of the target language.

23 languages supported

We support 23 languages: Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. If a language isn't on this list, it is not supported.

This list covers major global markets and specific regional needs, ensuring your application can reach a wide audience without language barriers.

Any voice speaks any language

You are not limited to matching voice gender or origin to the language. Any of the eight presets can produce any of the 23 supported languages. This allows you to pick the voice that best fits your brand's personality, whether it is a calm instructional tone or an energetic promo style, for any language.

This flexibility simplifies your integration logic, as you do not need to manage voice-to-language mappings separately.

Auto-detection for Latin script

If you do not specify a language, the API guesses it from the script. Latin text defaults to English. For non-English Latin-script text, set the language parameter explicitly to ensure accuracy.

This automatic detection works for the supported languages. It does not support Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali, or other unlisted languages.

  • Set language explicitly for non-English Latin text.
  • Default is English for Latin script if not specified.
  • Not supported: Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali.

Limitations

Sound wave illustration for: Limitations

The API returns complete audio files rather than streaming partial audio chunks. This means you cannot play audio while it is still being generated, which is ideal for batch processing but less suitable for real-time conversational interfaces where low latency is critical. You must wait for the full response before playback can begin.

There is no support for SSML (Speech Synthesis Markup Language), so you cannot control pronunciation or pauses using XML tags. Additionally, the voices do not support emotion tags, meaning you cannot explicitly direct a voice to sound happy or sad. You must rely on the natural intonation of the preset voices and adjust speed to convey tone.

No streaming or WebSocket

The API returns complete audio files rather than streaming partial audio chunks. This means you cannot play audio while it is still being generated, which is ideal for batch processing but less suitable for real-time conversational interfaces where low latency is critical. You must wait for the full response before playback can begin.

There is no support for SSML (Speech Synthesis Markup Language), so you cannot control pronunciation or pauses using XML tags. Additionally, the voices do not support emotion tags, meaning you cannot explicitly direct a voice to sound happy or sad. You must rely on the natural intonation of the preset voices and adjust speed to convey tone.

No SSML support

The API returns complete audio files rather than streaming partial audio chunks. This means you cannot play audio while it is still being generated, which is ideal for batch processing but less suitable for real-time conversational interfaces where low latency is critical. You must wait for the full response before playback can begin.

There is no support for SSML (Speech Synthesis Markup Language), so you cannot control pronunciation or pauses using XML tags. Additionally, the voices do not support emotion tags, meaning you cannot explicitly direct a voice to sound happy or sad. You must rely on the natural intonation of the preset voices and adjust speed to convey tone.

No emotion tags

The API returns complete audio files rather than streaming partial audio chunks. This means you cannot play audio while it is still being generated, which is ideal for batch processing but less suitable for real-time conversational interfaces where low latency is critical. You must wait for the full response before playback can begin.

There is no support for SSML (Speech Synthesis Markup Language), so you cannot control pronunciation or pauses using XML tags. Additionally, the voices do not support emotion tags, meaning you cannot explicitly direct a voice to sound happy or sad. You must rely on the natural intonation of the preset voices and adjust speed to convey tone.

Developer Experience

Sound wave illustration for: Developer Experience

The Readly API is designed for indie developers who want to integrate TTS without managing complex infrastructure. By mirroring the OpenAI SDK request shape, you can leverage familiar tools and libraries, changing only the base URL to point to our API. This means your existing code structure often requires minimal adjustment, allowing you to focus on building your product rather than debugging integration quirks.

We prioritize clarity in every interaction. Whether you are using plain HTTP or an SDK, you receive straightforward JSON responses that clearly indicate success or failure, helping you handle errors gracefully in your SaaS notifications and interfaces.

Simple curl examples

Getting started is as simple as pasting a single command into your terminal. Our documentation provides ready-to-use curl examples that demonstrate a complete request cycle, from setting the Authorization header to defining the voice and speed parameters. This allows you to test the API immediately without writing a single line of application code first.

You can copy these examples directly into your scripts, ensuring that your integration logic is correct before you invest time in building the full workflow into your application.

  • Copy-paste ready commands for quick testing
  • Adjustable parameters for voice, format, and speed
  • Works with standard HTTP clients

Clear error messages

When something goes wrong, you will know exactly why. Our API returns descriptive HTTP status codes and JSON error bodies that explain issues like rate limits, invalid character counts, or authentication failures. This transparency helps you debug integration problems quickly, even when handling automated batch requests.

You do not need to guess the cause of a failure; the response tells you if the input text was too long, the voice ID was invalid, or the key was expired, allowing for immediate corrective action.

  • Descriptive HTTP status codes
  • JSON error bodies with specific reasons
  • Immediate feedback on input validation

Dashboard for usage tracking

Your /get-api-key/ page serves as your command center. Here you can view your current balance, track characters consumed, and monitor your usage history in real-time. This transparency ensures you always know how much credit you have left and how your API usage correlates with your application's activity.

The dashboard also displays your active API key, allowing you to revoke and regenerate it instantly if needed, giving you full control over your integration's security without waiting for support.

  • Real-time balance and character tracking
  • Usage history logs
  • Instant key management

Comparison to Other Solutions

Sound wave illustration for: Comparison to Other Solutions

Full TTS suites often bundle speech-to-text, translation, and dubbing pipelines, requiring complex infrastructure and higher minimum spend. Readly API focuses strictly on text-to-speech, delivering just the audio file you need without unnecessary overhead or enterprise contract layers.

Simpler than full TTS suites

We provide a single endpoint for synthesis, avoiding the complexity of managing multiple microservices. You send text and receive an audio file, with no need to configure separate components for voice selection or format conversion.

Lower cost than enterprise plans

Enterprise plans often charge per seat or month, regardless of usage. We charge $0.01 per 1,000 characters with no monthly subscriptions. Failed requests are free, and you only pay for what you actually synthesize.

OpenAI SDK compatibility

Our API shape matches the OpenAI speech endpoint. You can use official OpenAI SDKs by simply changing the base_url to https://api.playpx9.com/v1. This allows indie developers to swap in high-quality TTS with minimal code changes.

Getting Started

Sound wave illustration for: Getting Started

Start with zero risk. Sign up using your Google account to receive $0.50 in free trial credit, valid for 7 days. This gives you about 50,000 characters to test the API without entering payment details or committing to a plan. No email/password accounts start with trial credit.

Once registered, your dashboard displays your unique API key, current balance, and usage statistics immediately. You can generate a new key at any time, which invalidates the old one instantly, ensuring security without downtime. The interface provides a ready-to-use curl example to help you begin.

Sign up for free trial

Create an account on the site's /get-api-key/ page using Google for instant access. You receive $0.50 of credit (approx. 50,000 characters) valid for 7 days. This allows you to test quality and integration without financial commitment.

Email-and-password accounts do not receive trial credit. There are no monthly plans, seats, or minimum spends required to start testing.

  • Instant access with Google sign-in
  • $0.50 credit (50,000 characters)
  • Valid for 7 days
  • No credit card required

Get API key instantly

Your API key appears immediately on the dashboard. Use the provided curl example or the OpenAI SDK to send a POST request to the speech endpoint. You will receive an audio file in your chosen format, though short phrases may take a few seconds due to overhead.

Manage your key directly from the interface. A new key issued replaces the old one at once, ensuring secure access for your indie projects.

  • Key displayed on dashboard
  • Compatible with OpenAI SDKs
  • Instant key rotation
  • Ready curl examples

Make first request in minutes

Send a simple POST request with your text, voice, and format preferences. Receive the audio file directly in the response. For longer texts, split input into 4,096-character chunks and join the resulting audio files.

The API processes requests sequentially. Under load, a request may wait in a queue, but typical speed is roughly real-time for long texts.

  • Simple JSON POST request
  • Direct audio file response
  • Supports 4096 char chunks
  • Processes sequentially

Support and Contact

Sound wave illustration for: Support and Contact

We keep things simple so you can focus on building your product. All support is handled via email, ensuring every query is documented and addressed thoroughly without the noise of phone queues.

Email support only

Our team monitors a single inbox to provide consistent, written responses. You can reach us at [email protected] for technical questions, billing issues, or general inquiries.

No phone number available

There is no phone line to call. This allows us to maintain asynchronous communication, which is often more efficient for developers who prefer ticket-style tracking over real-time voice calls.

Documentation available online

For immediate answers, consult our comprehensive documentation at /docs/. It covers integration steps, error codes, and code examples, often resolving issues faster than waiting for an email reply.

FAQs

Sound wave illustration for: FAQs

The Readly API is built for reliability, processing requests sequentially with transparent queueing during high load. While we don’t offer an SLA, our infrastructure handles standard SaaS notification volumes efficiently, ensuring your audio files are generated and returned without unexpected downtime.

You don’t need a dedicated DevOps team to use it. The API follows standard REST conventions, meaning you can integrate it using simple HTTP POST requests or even the official OpenAI SDKs with a minor base_url change. It’s designed for indie developers who want to ship features quickly without managing complex audio pipelines.

Is the API reliable?

We process requests sequentially, so under heavy load, your request may wait in a queue rather than failing. This ensures stable output but means latency increases slightly during peak times. We don’t offer an SLA, but our system is designed for consistent, predictable delivery for SaaS notifications.

Do I need a developer?

A basic understanding of APIs is helpful, but our endpoints are straightforward. You can generate audio with a single POST request. Our documentation provides ready-to-use curl examples, making it easy for any developer to integrate the service into their dashboard or notification system.

Can I use it commercially?

Yes. The API is fully commercial-ready. You pay only for what you use, with no minimum spend or enterprise lock-in. Generated audio files can be used in your SaaS products, apps, or notifications without additional royalties, provided you adhere to our terms of service.

Is data private?

Your input text is used solely to generate the audio and is not used for training our models. The generated audio is returned in the response and not stored on our servers. We retain request metadata (time, characters, status) strictly for billing and usage tracking purposes.

What if I make a mistake?

If a synthesis request fails due to a technical error, you are not charged for that request. The credit is returned to your balance automatically. You can retry the request with corrected parameters. For other issues, such as incorrect voice selection, no charge is applied.

How do I update my key?

Can I use multiple languages?

The API supports 23 languages, including English, Spanish, French, German, Japanese, and Arabic. Any voice can speak any of these languages. If you don’t specify a language, the system guesses it from the script, but we recommend setting the language field explicitly for non-English Latin text.

Is the voice realistic?

Does it work on mobile?

Yes. The API returns standard audio files (MP3, WAV, OPUS, etc.) that can be played on any mobile device. You can integrate the API into your mobile app or use it to generate audio assets for your mobile interface. There is no specific mobile SDK required.

Can I customize the voice?

What is the watermark?

Is training data used?

Can I cancel anytime?

Do I need a contract?

Listen to the 8 voices

Each sample below was generated by this API from the same sentence.

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. OpenAI API reference: create speech
  3. W3C WAI: making audio and video accessible
  4. FFmpeg documentation
  5. MDN: Web Speech API
  6. openai-python on GitHub
  7. LibriTTS-R paper (arXiv 2305.18802)
  8. Node.js documentation
  9. LibriTTS-R corpus (OpenSLR 141)
  10. Python documentation
  11. openai-node on GitHub
  12. Creative Commons Attribution 4.0

More on this site