Speakable Cloud TTS API for Scalable Voice Solutions

By Speakable editorial team · Updated 2026-11-19

This page compares the browser-native Chrome Web Speech API with our cloud TTS API to help indie developers choose the right audio solution for their apps. We analyze latency, voice quality, and integration complexity using concrete data, so you can decide whether to stick with built-in browser voices or switch to a consistent, high-quality cloud service for scalable features.

Written for startup engineers and indie hackers, this guide covers real-world workflows, pricing models, and technical constraints. You will find clear distinctions between short UI feedback and long-form content generation, along with specific details on our 8 voice presets, 23 languages, and per-character billing to avoid vendor lock-in.

Speakable Cloud TTS API for Scalable Voice Solutions

Frequently asked questions

Is the Chrome Web Speech API free?

The Chrome Web Speech API is free and built into the browser, making it ideal for quick UI interactions. However, it relies on the user’s device capabilities, which can lead to inconsistent voice quality and limited language support across different platforms and operating systems.

How does cloud TTS pricing compare?

Speakable charges $10 per 1 million characters, or $0.01 per 1,000 characters. This predictable, usage-based model avoids monthly subscriptions or minimum spend requirements, allowing indie developers to scale costs directly with audio generation volume without upfront commitments or hidden fees.

Can I use the cloud API with mobile apps?

Yes. The API is a standard HTTP endpoint compatible with any environment that can send JSON requests. Since it uses the same request shape as the OpenAI speech endpoint, you can integrate it into mobile apps by adjusting the base URL in your SDK.

What languages does the cloud API support?

The API supports 23 languages, including English, Spanish, French, German, Japanese, and Arabic. Any voice can speak any supported language. If the language parameter is omitted, the system guesses the language based on the script, defaulting to English for Latin text.

Do cloud TTS voices sound robotic?

Voices are based on licensed recordings from LibriTTS-R and LibriVox, providing a natural, human-like quality. While generated audio carries an inaudible watermark to mark it as synthetic, the voices are designed to be clear and expressive, avoiding the typical robotic sound of older TTS systems.

How long can a text request be?

Each request supports up to 4,096 characters of input text. For longer texts, you should split the content into parts of this size and join the resulting audio files. This ensures consistent generation quality and avoids timeouts during the synthesis process.

Can I change voices mid-sentence?

No, the API does not support changing voices mid-sentence within a single request. Each request generates audio for one voice. To switch voices, you must send separate requests for each segment of text and concatenate the resulting audio files in your application.

Is there a free trial for the API?

Yes, signing in with Google provides $0.50 of credit, valid for 7 days, which covers approximately 50,000 characters. Accounts created with email and password start without this trial credit. You can top up later using USDT on the TRON network.

How do I authenticate my requests?

Include your API key in the Authorization header as a Bearer token: Authorization: Bearer tts-.... Each account has one active key. You can issue a new key at any time, which immediately invalidates the old one, ensuring secure access to the endpoint.

What happens if a request fails?

If synthesis fails, the charge is returned in full. You are only billed for successful requests. This ensures you do not pay for audio that was not generated, protecting your budget while you experiment with different voices, speeds, or formats.

Can I clone my own voice?

No, the API does not offer custom voice cloning or allow you to upload your own voice. It provides eight preset voices based on licensed recordings. If you need a specific voice, you must choose from the available presets like Aria, Nova, or Felix.

Does the API support SSML?

No, the API does not support SSML (Speech Synthesis Markup Language) or phoneme dictionaries. You control pacing and emphasis primarily through the speed parameter, which allows values between 0.25 and 4.0, or by structuring your input text with appropriate punctuation and line breaks.

Is there a limit on requests per minute?

Yes, the limit is 60 requests per minute per key. Before your first top-up, the limit is 10 requests per minute. If you exceed the limit, requests may wait in a queue or be rejected. Plan your batching strategy accordingly for high-volume applications.

Do I need a credit card to start?

No credit card is required. You can sign up with Google for a free trial credit. Payments are made via USDT on the TRON network (TRC-20), with a minimum top-up of $5. There are no monthly plans, subscriptions, or invoices with VAT.

First request

curl

curl https://api.s9999.info/v1/audio/speech \
  -H "Authorization: Bearer $TTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"tts-1","voice":"aria","input":"Hello from the API.","response_format":"mp3"}' \
  --output hello.mp3

The Browser-Native Option

Sound wave illustration for: The Browser-Native Option

The Chrome Web Speech API is built into every modern browser, allowing indie developers to add text-to-speech to their UI without managing a backend service. This native approach eliminates the need for API keys or network requests, making it ideal for immediate, low-latency interactions like button announcements or simple navigation cues where absolute audio fidelity is secondary to speed.

Zero latency for short phrases

Zero latency for short phrases

Since the speech engine runs locally on the user's device, there is no network round-trip time. This results in instantaneous playback for brief UI strings, ensuring that interactive feedback feels seamless and responsive without waiting for a server to generate audio data.

  • Instant playback for short strings
  • No network dependency for basic phrases
  • Works offline once the page loads

Free with no API key needed

There are no usage costs, rate limits, or billing cycles to configure. You do not need to sign up for a service, manage credentials, or worry about character counts. This makes it a zero-friction starting point for prototypes or apps with minimal audio requirements.

  • No API key required
  • Zero cost per request
  • No account setup needed

Limited voice selection

The available voices depend entirely on the user's operating system and browser. You cannot guarantee a specific voice or language across all devices, and the range is often restricted to basic male or female presets without the nuanced tones or multi-language support found in dedicated cloud APIs.

  • Dependent on OS and browser
  • Limited voice variety
  • No consistent quality across devices

The Cloud API Alternative

Sound wave illustration for: The Cloud API Alternative

The cloud API delivers consistent, high-quality audio generation independent of the user’s browser or device capabilities. Unlike native APIs that vary by OS, our service ensures uniform output for every request, making it ideal for apps that require reliable audio delivery across diverse environments.

With dedicated infrastructure, you avoid the unpredictability of local synthesis.

Consistent audio quality

Every generated file maintains a standard of clarity and tone, avoiding the robotic or inconsistent artifacts common in browser-based synthesis. This reliability allows you to build audio features that feel polished from day one, without worrying about edge cases in specific browser versions.

The engine processes requests sequentially, providing stable results. While short phrases may take a few seconds, the output quality remains steady, ensuring that your app’s audio features do not degrade under varying network conditions.

  • Uniform output across all devices
  • No browser-specific artifacts
  • Predictable generation latency

8 distinct voice presets

Choose from eight curated voices, including warm narrators like Aria and calm instructors like Sage. Each voice is based on licensed recordings from LibriVox, offering a professional tone that fits various content types, from news to storytelling.

The presets cover a range of tones, from energetic promos with Felix to deep documentaries with Orion. This variety allows you to match the audio to your brand’s voice without needing custom voice design or cloning services.

  • Aria, Nova, Luna, Sage
  • Felix, Milo, Atlas, Orion
  • Licensed LibriVox recordings

23 supported languages

Support major global languages including English, Spanish, German, and Japanese. Any voice can speak any of these languages, allowing you to localize your app without managing multiple voice models. The API also guesses the language from the script if not specified.

For non-English Latin-script text, you can explicitly set the language parameter to ensure accurate pronunciation. This flexibility covers most key markets, though languages like Vietnamese and Thai are not yet supported.

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

When to Use Each

Sound wave illustration for: When to Use Each

The Chrome Web Speech API offers near-zero latency for immediate UI interactions, making it ideal for brief system notifications or button clicks where milliseconds matter. It requires no API key and incurs no cost, allowing you to prototype quickly without infrastructure overhead. However, voice quality varies significantly across different browsers and operating systems, which can lead to inconsistent user experiences as your app scales.

For longer-form content, a cloud TTS API provides consistent, high-quality audio regardless of the user's device. With 8 distinct voice presets and support for 23 languages, you can maintain a uniform brand voice across all platforms.

UI feedback and notifications

Use the native browser API for transient sounds like click confirmations or error alerts. The instant response time ensures the user feels immediate feedback without waiting for a network request.

Reserve the cloud API for persistent audio elements or background narration where consistency is key. If your app relies on audio for navigation or critical information delivery, the standardized voices and formats of a cloud API prevent the jarring effect of switching browser voices mid-session.

  • Ideal for: Button clicks, form validation sounds, short alerts
  • Best for: Background narration, consistent character voices, multi-language support

Content generation and podcasts

When generating longer texts, such as blog summaries or podcast episodes, the cloud API’s ability to handle up to 4096 characters per request ensures high-fidelity output. You can split longer texts into parts and join the audio files, maintaining a seamless listening experience. The variety of voices allows for dynamic storytelling that native browser voices often cannot replicate.

Note that generation is not real-time streaming; requests are processed sequentially. For a 1-minute podcast clip, expect a few seconds of processing time. This delay is acceptable for on-demand content but would be frustrating for live interactive scenarios. Use the API when quality trumps instant availability.

Multi-user production environments

In multi-user settings, the cloud API’s rate limits (60 requests per minute per key) and predictable pricing ($0.01 per 1,000 characters) allow for better capacity planning. Unlike the browser API, where quality depends on the user’s device, the cloud API guarantees the same voice and format for every user, ensuring brand consistency at scale.

Track your usage via the dashboard to monitor character consumption and balance. Since failed requests are free, you can implement retry logic without fear of unexpected costs. This predictability makes the cloud API suitable for production environments where audio quality is a core feature rather than an experimental add-on.

Technical Comparison

Sound wave illustration for: Technical Comparison

The Chrome Web Speech API operates entirely within the browser, requiring no network calls or API keys. Our cloud TTS API uses standard HTTP POST requests to https://api.s9999.info/v1/audio/speech, accepting JSON bodies with model, input, and voice fields. This structure allows you to leverage existing OpenAI SDKs by simply adjusting the base_url.

Request formats and endpoints

Send text via JSON to the speech endpoint. The API returns audio files in formats like MP3, WAV, OPUS, FLAC, or PCM. Authentication is handled via a Bearer token in the Authorization header, ensuring secure access to your account's credits.

  • POST /v1/audio/speech for generation
  • GET /v1/voices for metadata
  • GET /v1/models for available engines

Character limits per request

Each request accepts up to 4,096 characters of text. For longer content, split your input into chunks of this size and join the resulting audio files. This approach maintains consistent quality while handling extensive scripts or articles efficiently.

Rate limiting strategies

The API allows 60 requests per minute per key. Before your first top-up, this limit drops to 10 requests per minute. Under load, requests are processed sequentially, so longer texts may queue briefly. Monitor your usage via the dashboard to avoid throttling.

  • 60 requests/minute after top-up
  • 10 requests/minute initially
  • Sequential processing under load

Voice Quality and Variety

Sound wave illustration for: Voice Quality and Variety

The eight available voices are designed for specific use cases, ranging from warm narrators to energetic promotional speakers.

Whether you need a calm, instructional tone for tutorials or a deep, documentary-style voice for podcasts, there is a preset that matches your app's aesthetic.

Narrator vs conversational tones

Aria offers a warm, clear narration ideal for audiobooks or explanatory content. In contrast, Felix provides an energetic, promo-ready voice perfect for advertisements or quick notifications. Milo delivers a friendly, conversational tone that feels natural for interactive dialogues.

Luna’s soft storytelling style suits immersive experiences, while Sage’s calm instructional voice works well for guided tours. Atlas maintains a neutral news tone, and Orion provides a deep, documentary-style delivery for serious content. These distinct personalities allow you to match audio to your brand’s voice.

Gender and tone variety

The API offers four female voices (Aria, Nova, Luna, Sage) and four male voices (Felix, Milo, Atlas, Orion), providing balanced options for diverse applications.

Each voice is based on licensed recordings. The distinct characteristics allow for clear differentiation between different sections of your app, enhancing user experience through auditory variety without requiring additional configuration.

  • Aria: Warm, clear narrator
  • Nova: Bright, upbeat
  • Luna: Soft, storytelling
  • Sage: Calm, instructional
  • Felix: Energetic, promo
  • Milo: Friendly, conversational
  • Atlas: Neutral, news
  • Orion: Deep, documentary

Language support coverage

Speakable 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. Any voice can speak any of these languages, offering flexibility in global deployments.

If the language field is not set, the system guesses the language from the script, defaulting to English for Latin text. For non-English Latin-script text, it is recommended to explicitly set the language parameter to ensure accurate pronunciation and tone.

Voice Quality and Variety

Pricing Models

Sound wave illustration for: Pricing Models

Speakable charges $10 per million characters, effectively $0.01 per 1,000 characters. Spaces and punctuation count toward this total, ensuring predictable costs for your audio infrastructure. Failed synthesis requests are free, so you only pay for successful audio generation.

Balances never expire, and there are no monthly subscriptions or minimum spend requirements.

Per-character billing

Speakable charges $10 per million characters, effectively $0.01 per 1,000 characters. Spaces and punctuation count toward this total, ensuring predictable costs for your audio infrastructure. Failed synthesis requests are free, so you only pay for successful audio generation.

Free trial credits

New accounts signing in with Google receive $0.50 in trial credit, valid for 7 days. This provides approximately 50,000 characters for testing. Email-based accounts start without this initial credit.

USDT top-up flexibility

Top-ups use USDT on the TRON network (TRC-20) with a $5 minimum. Credits are automatically applied within minutes of transfer. We do not process credit cards, PayPal, or offer traditional invoicing with VAT.

  • Minimum top-up: $5
  • Network: TRON (TRC-20)
  • Processing time: ~minutes
  • No monthly plans

Audio Formats

Sound wave illustration for: Audio Formats

Choose MP3 for standard web playback to balance file size and compatibility. WAV provides uncompressed high fidelity for local processing or archival. FLAC and OPUS offer lossless and efficient lossy alternatives for specific technical needs.

All formats are returned directly in the response, allowing immediate use without additional conversion steps.

MP3 for web playback

MP3 is a common choice for web applications. It provides a good balance between audio quality and file size, ensuring fast loading times for users.

This format works seamlessly with standard HTML5 audio players and is supported by all modern browsers.

  • Best for general web playback
  • Small file size
  • Universal browser support

WAV for high fidelity

WAV delivers uncompressed, high-fidelity audio. Use this format when preserving the exact quality of the generated speech is critical, such as for professional presentations or archival purposes.

Note that WAV files are significantly larger than MP3, which may impact bandwidth usage for high-volume applications.

  • Uncompressed audio
  • Highest quality
  • Larger file size

FLAC and OPUS options

FLAC offers lossless compression, preserving all audio data while reducing file size compared to WAV. OPUS provides efficient lossy compression, ideal for low-bandwidth environments or mobile applications.

These formats give you flexibility depending on your specific storage and bandwidth constraints without sacrificing the core audio integrity.

  • FLAC: Lossless compression
  • OPUS: Efficient lossy compression
  • Ideal for mobile/low-bandwidth

Integration Complexity

Sound wave illustration for: Integration Complexity

Speakable’s API mirrors the OpenAI speech endpoint structure, allowing indie developers to reuse existing codebases with minimal changes. By simply updating the base_url to https://api.s9999.info/v1, teams can leverage official OpenAI SDKs for seamless integration.

OpenAI SDK compatibility

You can integrate Speakable using the standard OpenAI libraries by pointing them to our API base.

  • Change base_url to https://api.s9999.info/v1
  • Use standard POST request shapes
  • Reuse existing OpenAI SDK client logic

Simple curl requests

For quick testing or lightweight implementations, a single curl command suffices. The JSON body requires only model, input, and voice, making it easy to prototype without setting up a full SDK environment.

  • Supports standard JSON payloads
  • Requires no complex headers beyond Bearer token
  • Ideal for rapid prototyping

Streaming vs batch processing

Unlike some competitors, Speakable does not offer WebSocket or real-time streaming. It operates on a batch model: you send text, wait for processing, and receive the final audio file. This simplifies error handling and reduces complexity for indie apps.

  • No streaming support
  • Batch processing model
  • Simpler error handling

Data Privacy and Usage

Sound wave illustration for: Data Privacy and Usage

We prioritize transparency in how your data moves through our pipeline. Your input text is processed strictly to generate the requested audio file and is not retained for long-term storage after the response is delivered. We keep only the necessary metadata—such as request time, character count, and status—specifically for accurate billing and usage tracking.

Input text storage policies

Generated audio is returned directly in the response and is not stored on our servers. This ensures that your specific content remains ephemeral unless you choose to save it yourself.

Training data usage

Unlike many consumer-grade TTS services, we do not use your input text to train our models. Your data is used solely for the immediate synthesis task, ensuring your proprietary content remains yours.

Synthetic audio watermarks

Every audio file generated through our API carries an inaudible watermark. This marks the content as synthetic speech, helping downstream platforms identify AI-generated audio while remaining transparent to your end-users.

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. This sequential processing guarantees that each audio file is generated correctly without race conditions.

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. TRON developer hub: TRC-20
  3. openai-python on GitHub
  4. openai-node on GitHub
  5. W3C WAI: making audio and video accessible
  6. ISO 639 language codes
  7. LibriVox public-domain audiobooks
  8. LibriTTS-R paper (arXiv 2305.18802)
  9. Python documentation
  10. Tether (USDT)
  11. Creative Commons Attribution 4.0
  12. OpenAI API reference: create speech

More on this site