Cloud TTS API: A Practical Guide for Indie Developers

By Speakable editorial team · Updated 2026-11-19

This guide helps indie developers and startups choose between the Chrome Web Speech API for quick UI interactions and a cloud TTS API for consistent, high-quality audio at scale. We provide honest comparisons, clear pricing, and technical details for building reliable speech features without vendor lock-in.

You will find practical insights into voice quality, integration ease, and data privacy. Learn how to handle rate limits, select from eight presets, and manage billing with transparent per-character costs. This resource is designed for teams that value specific, factual data over marketing hype.

Cloud TTS API: A Practical Guide for Indie Developers

Frequently asked questions

Which TTS API is best for startups?

Speakable is designed for indie developers who need consistent quality without vendor lock-in. We offer a transparent API compatible with OpenAI SDKs, simple pricing at $10 per million characters, and no monthly plans. This structure helps startups control costs while accessing high-quality synthetic speech for their applications.

How do voices compare across APIs?

We provide eight distinct presets like Aria, Nova, and Atlas, based on licensed recordings from LibriTTS-R and LibriVox. Each voice carries an inaudible watermark marking it as synthetic. While we do not offer custom cloning, our voices cover 23 languages with natural, clear delivery suitable for various content types.

Is the cloud API cheaper than others?

Our rate is $10 per million characters, or $0.01 per 1,000 characters, with failed requests free. We do not offer monthly plans or subscriptions, allowing you to pay only for what you use. This model suits startups with variable traffic, avoiding the overhead of tiered enterprise contracts.

Which API has the most languages?

Our API supports 23 languages, including Arabic, Chinese, English, French, German, Japanese, and Spanish. Any voice can speak any of these languages. We do not support Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, or Bengali. You must specify the language ISO code for non-English Latin-script text to ensure accuracy.

Can I use the API for commercial projects?

Yes, you can use generated audio for commercial projects. The input text is used only to produce the audio and is not used for training. Generated audio is returned in the response and not stored, ensuring your data remains yours while you build your product with synthetic speech.

Is there a free trial?

Yes, signing in with Google provides $0.50 of credit, valid for 7 days, allowing about 50,000 characters of generation. Accounts created with email and password start without trial credit. You can top up with USDT on the TRON network, with no minimum spend or expiration on your balance.

How easy is it to integrate?

Integration is straightforward using standard HTTP requests or the OpenAI SDK by changing the base URL. Send a POST request to /v1/audio/speech with your text, voice, and format. We provide a ready curl example on the dashboard, making it easy for developers to start generating audio immediately.

Does the API support streaming?

No, we do not support streaming, WebSocket connections, or real-time partial audio. You send text and receive the complete audio file. For long texts, you must split the input into parts of up to 4096 characters and join the resulting audio files manually in your application logic.

Can I clone voices?

No, we do not offer custom voice cloning or the ability to upload your own voice. We provide eight fixed presets like Felix, Milo, and Orion. These voices are based on licensed recordings and are available for immediate use without additional configuration or training periods for custom samples.

Is the audio quality good?

Our voices are based on licensed recordings from LibriTTS-R and LibriVox, ensuring clear and natural speech. Each file includes an inaudible watermark identifying it as synthetic. The engine supports two models, tts-1 and tts-1-hd, both offering high-quality output suitable for professional applications without noticeable artifacts or robotic tones.

What formats are supported?

We support mp3, wav, opus, flac, and pcm formats. You can specify the desired format in your request body. This flexibility allows you to choose the best balance between file size and quality for your specific use case, whether for web playback, storage, or further audio processing.

Are there rate limits?

The limit is 60 requests per minute per key, or 10 per minute before your first top-up. Requests are processed sequentially, so under load, you may experience queuing. Longer texts take time to generate, roughly real-time speed, so plan your requests accordingly to avoid delays in your user workflow.

How is data handled?

Input text is used only to produce the audio and is not used for training. Generated audio is returned in the response and not stored. We keep request metadata, such as time, characters, and status, solely for billing purposes. Your data remains private and is not shared with third parties.

Who is the API best for?

Speakable is built for indie developers and startups needing consistent, high-quality text-to-speech without complex enterprise contracts. Our API offers simple pricing, no vendor lock-in, and easy integration via standard HTTP requests. This straightforward approach makes it a practical choice for building reliable audio features in modern web applications, from quick UI interactions to scalable production workloads.

Selection Criteria

Sound wave illustration for: Selection Criteria

Choosing a text-to-speech engine requires balancing audio fidelity with operational simplicity. For indie developers, the decision often hinges on whether the output sounds natural enough for end-users and whether the integration path is straightforward enough to maintain within a small engineering team.

Cost structure is equally critical. Transparent per-character billing prevents budget surprises during scale-up, while clear rate limits ensure that sudden traffic spikes do not silently degrade the user experience or incur unexpected charges.

Voice quality and variety

The engine offers eight distinct presets, ranging from warm narrators to energetic promo voices. This variety allows developers to match audio to specific content types without needing complex configuration or custom voice design.

Each voice is generated from licensed recordings, ensuring consistent quality across all outputs. The selection covers different tones, from soft storytelling to neutral news delivery, providing enough variety for most startup applications.

  • 8 distinct voice presets
  • Licensed source recordings
  • Consistent audio quality

Pricing transparency

Billing is strictly per character at $0.01 per 1,000 characters. Spaces and punctuation count, so input length directly correlates with cost. Failed requests are free, ensuring you only pay for successful synthesis.

There are no monthly plans, seats, or minimum spend requirements. A $0.50 free trial provides initial credit for testing, and top-ups via USDT on the TRON network (TRC-20) are credited automatically without hidden fees or subscription locks.

Ease of integration

The API follows the standard OpenAI request shape, meaning existing OpenAI SDKs work by simply changing the base URL. This reduces development time for teams already familiar with common TTS patterns.

Authentication uses a simple Bearer token, and endpoints are standard HTTP POST/GET requests. Developers can start with a basic curl example and scale up by splitting long texts into 4,096-character chunks.

Top Contender: Cloud TTS API

Sound wave illustration for: Top Contender: Cloud TTS API

The Speakable API distinguishes itself through a curated selection of eight voice presets, each engineered for specific tonal needs. From the warm clarity of Aria to the deep documentary style of Orion, developers can match audio output precisely to their content’s mood without external tools.

Language support extends to 23 distinct languages, ensuring broad accessibility. The system defaults Latin script text to English, but allows explicit language setting for accurate pronunciation in non-English contexts, reducing ambiguity for international user bases.

8 voice presets

Each voice is based on licensed recordings from LibriTTS-R, providing a consistent, high-quality baseline. The eight presets cover a spectrum from upbeat narration to calm instruction, allowing indie developers to maintain brand consistency across different content types.

  • Female: Aria, Nova, Luna, Sage
  • Male: Felix, Milo, Atlas, Orion

23 languages supported

Support includes major global languages such as English, Spanish, German, Japanese, and Hindi. The API handles script detection intelligently, defaulting to English for Latin text unless a specific language code is provided.

$0.01 per 1000 characters

Pricing is straightforward: $10 per million characters, or $0.01 per 1,000 characters. Spaces and punctuation count toward the total, ensuring predictable costs. Failed requests incur no charge, protecting your budget during development and testing phases.

Voice Options

Sound wave illustration for: Voice Options

Speakable offers eight distinct voice presets derived from licensed recordings and public-domain audiobook readings. These voices provide a versatile palette for indie developers, ensuring your audio content matches the specific tone of your application without requiring custom voice design.

Female voices: Aria, Nova, Luna, Sage

Aria delivers a warm, clear narration style ideal for general content. Nova offers a bright, upbeat tone suited for energetic interfaces. Luna provides a soft, storytelling voice perfect for narratives, while Sage presents a calm, instructional tone for educational materials.

  • Aria: warm, clear narrator
  • Nova: bright, upbeat
  • Luna: soft, storytelling
  • Sage: calm, instructional

Male voices: Felix, Milo, Atlas, Orion

Felix brings an energetic, promo-ready quality to your projects. Milo offers a friendly, conversational style for user interactions. Atlas provides a neutral, news-like delivery, and Orion adds a deep, documentary tone for serious content.

  • Felix: energetic, promo
  • Milo: friendly, conversational
  • Atlas: neutral, news
  • Orion: deep, documentary

Tone and style variety

Each generated file carries an inaudible watermark marking it as synthetic speech. The eight presets cover a wide range of emotional tones, allowing you to select the exact voice that fits your brand’s personality.

Language Support

Sound wave illustration for: Language Support

Language support is a critical factor for indie developers building global products. Speakable’s API covers 23 languages, ensuring your application can reach a diverse audience without needing multiple specialized providers. This broad coverage simplifies your tech stack while maintaining high audio quality across different regions.

Whether you are localizing a podcast or creating e-learning modules, having reliable voice synthesis in multiple languages reduces development complexity. You can focus on content rather than managing disparate language engines.

Broad coverage with 23 languages

The API supports 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 selection covers major global markets and many emerging regions, allowing you to serve a wide user base with consistent voice quality.

Each voice can speak any of these languages, so you do not need to assign specific voices to specific languages. This flexibility lets you maintain brand consistency by using the same voice character across different language versions of your app.

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

Auto-detection of Latin script

If you do not specify a language, the API guesses the language from the script. Latin script text defaults to English, which is convenient for most English-speaking developers.

This auto-detection works for other scripts as well, reducing the need for explicit configuration in simple use cases. However, relying on detection can lead to inaccuracies if the text contains mixed languages or ambiguous phrasing.

Explicit language setting

For non-English Latin-script text, you should explicitly set the language field in your request. This ensures the voice pronounces words correctly and avoids misinterpretation by the engine.

Using the ISO code for the language parameter guarantees consistent results, especially for languages with complex phonetic rules or when your input text contains mixed scripts. This is essential for professional-grade localization.

  • Set language for non-English Latin text
  • Use ISO codes for precision
  • Avoids default English mispronunciation

Pricing Comparison

Sound wave illustration for: Pricing Comparison

Choosing a TTS provider often means navigating complex tiered plans or per-minute audio charges. Speakable simplifies this with a straightforward per-character billing model. You pay $0.01 per 1,000 characters, including spaces and punctuation. This predictable unit cost helps indie developers forecast expenses accurately without guessing how voice length or format affects the final invoice.

Unlike services that charge per audio second, character-based pricing aligns directly with your input. Whether you generate a short UI alert or a long narrative, the cost scales linearly. This transparency eliminates surprise overages, making it easier to integrate audio features into startup budgets without financial guesswork.

Per-character billing model

Speakable charges $0.01 per 1,000 characters. Spaces and punctuation count toward this total. This linear model ensures costs match your actual usage precisely, avoiding the ambiguity of per-minute audio billing where silence or pauses might inflate prices unfairly.

  • $0.01 per 1,000 characters
  • Spaces and punctuation count
  • Linear cost scaling

Free trial availability

New accounts can start with a $0.50 trial credit, valid for 7 days. This provides approximately 50,000 characters for testing voices and formats without requiring immediate payment details. It allows developers to validate quality before committing to larger volumes.

No hidden fees

There are no monthly subscriptions, seat fees, or minimum spend requirements. Failed requests are refunded in full, ensuring you only pay for successful synthesis. Top-ups via USDT on the TRON network are processed automatically, and your balance never expires. Simple, transparent, and scalable.

  • No monthly plans
  • No seat fees
  • Failed requests refunded
  • Balance never expires

Integration Ease

Sound wave illustration for: Integration Ease

Integrating Speakable into your indie stack is straightforward because the API mirrors the OpenAI speech endpoint structure. This means you can leverage existing OpenAI SDKs simply by updating the base URL to https://api.s9999.info/v1. This compatibility reduces boilerplate code and accelerates development for teams already familiar with standard TTS patterns.

The endpoint accepts standard JSON payloads, making it easy to integrate with any backend language. You send a POST request with your text, selected voice, and desired format, receiving a direct audio file response. This simplicity ensures minimal friction when adding voice features to apps or websites.

OpenAI SDK compatibility

By adopting the same request shape as the OpenAI speech endpoint, Speakable allows developers to use official OpenAI SDKs without rewriting client logic. You only need to change the base_url to https://api.s9999.info/v1 and provide your Speakable API key. This approach minimizes integration effort and leverages familiar authentication patterns.

  • Use official OpenAI SDKs with a simple URL swap
  • Familiar JSON request structure reduces learning curve
  • Standard Bearer token authentication with tts-... keys

Simple curl examples

For quick testing or scripting, the API supports direct HTTP requests. A basic curl command can generate speech by sending the model, input text, and voice parameters. This makes it easy to prototype audio features directly from the terminal or include in deployment scripts without setting up a full SDK environment.

  • Test integration instantly via command line
  • Ideal for backend scripts and automated workflows
  • No additional libraries required for basic usage

Standard HTTP methods

The API relies on standard RESTful conventions, primarily using POST for speech generation. This predictability means your existing error-handling and retry logic for HTTP requests works seamlessly. You can integrate Speakable into any microservice architecture or serverless function without needing custom protocol adapters.

  • Consistent POST endpoint for audio generation
  • Standard HTTP status codes for success and errors
  • Easy integration with existing API gateways

Audio Formats

Sound wave illustration for: Audio Formats

The API allows you to select the output format via the response_format field, giving you direct control over file size and compatibility. This flexibility ensures your audio assets are optimized for their specific delivery context, whether that is immediate web playback or archival storage.

MP3 for web use

MP3 is the default choice for most web applications due to its excellent compression and universal browser support. It provides a balance between audio clarity and bandwidth efficiency, making it ideal for dynamic UI elements or quick background narration in mobile apps.

WAV for high quality

WAV offers uncompressed audio, preserving the full fidelity of the synthesis engine. This format is suitable for professional editing workflows or high-fidelity playback scenarios where file size is less critical than audio integrity. It ensures no generation artifacts are introduced by compression.

FLAC and OPUS support

FLAC provides lossless compression, reducing file size without any quality loss, which is perfect for archival or podcast distribution. OPUS is highly efficient for low-bandwidth delivery or VoIP integration, offering superior quality at lower bitrates compared to MP3 for similar perceptual audio.

Audio Formats

Rate Limits and Scalability

Sound wave illustration for: Rate Limits and Scalability

Understanding rate limits is crucial for indie developers planning their API usage. Speakable’s cloud TTS API allows 60 requests per minute per key, enabling efficient batch processing for most startup workloads.

When traffic exceeds this limit, requests are processed in a queue. Under load, you might experience slight delays, but your audio files will be generated sequentially without data loss or duplication.

60 requests per minute

Each API key is limited to 60 requests per minute. New accounts start with 10 requests per minute until the first top-up is processed. This limit ensures stable service for all users.

You can monitor your usage through the dashboard, which tracks requests, characters, and balance in real time.

  • 60 requests/minute after top-up
  • 10 requests/minute initially
  • One key per account

Queue handling

Requests are processed one after another. If multiple requests are sent simultaneously, they wait in a queue. This sequential processing guarantees that each audio file is generated correctly without race conditions.

Typical speed is roughly real time: about 25 characters of text per second of generation. Short phrases take a few seconds, while longer texts may take minutes.

  • Sequential processing
  • No parallel generation
  • Predictable timing

Long text processing

For texts longer than 4,096 characters, split the input into parts and join the resulting audio files. This ensures compatibility with the API’s character limit per request.

Failed requests are free, so you can retry parts without worrying about extra charges. This makes error handling straightforward and cost-effective.

  • Max 4,096 characters/request
  • Split and join audio
  • Free retries on failure

Data Privacy

Sound wave illustration for: Data Privacy

We treat your input text as a transient payload, not a dataset. It is used solely to synthesize the requested audio file and is never consumed for training our models or improving voice engines.

No training on input text

Unlike many consumer-grade TTS providers, we do not collect or store the actual text content you send to the API. Once the audio is generated and returned in the response, the text payload is discarded and does not contribute to our training corpus.

No storage of generated audio

The resulting audio files are returned directly in the HTTP response and are not persisted on our servers after delivery. You retain full ownership and control over the generated files, ensuring your content remains private and secure without lingering copies on our infrastructure.

Metadata retention for billing

We retain minimal request metadata—specifically the timestamp, character count, and status code—to accurately calculate usage and process charges. This data is strictly limited to billing verification and is not used for analytics or user profiling, keeping your operational footprint transparent and lean.

Customer Support

Sound wave illustration for: Customer Support

Speakable prioritizes transparency and self-service to keep operational costs low for indie developers. Our support model is straightforward: we handle inquiries via email, ensuring every question receives a thoughtful, documented response without the pressure of immediate live interaction.

Email-only support

We respond to all support tickets via email. This allows our team to provide detailed, context-aware answers for technical integration issues or billing questions. There are no phone numbers to call or complex IVR systems to navigate; just send your query and wait for a direct reply.

  • Direct email access to support staff
  • Detailed technical responses for API issues
  • Clear billing explanations for USDT top-ups

No live chat

We do not offer live chat widgets. This decision reduces overhead and ensures that support interactions are searchable and accessible to all users, not just those online at a specific moment. You can submit questions anytime and receive a comprehensive answer without waiting in a queue.

Self-service docs

Most developers can resolve issues using our documentation. The /docs/ page covers all endpoints, voice presets, and integration examples. For quick testing, the /get-api-key/ page provides ready-to-use curl commands and real-time balance tracking, reducing the need to contact support for basic setup questions.

  • Complete API reference documentation
  • Curl examples for immediate testing
  • Real-time usage tracking on dashboard

Use Cases for Startups

Sound wave illustration for: Use Cases for Startups

Startups often need to produce audio content quickly without hiring voice actors. Our API allows you to generate podcast episodes or e-learning modules at scale, using realistic voices that maintain consistent quality across hundreds of files.

Podcast generation

You can feed scripts into the API to create full episodes. With 8 distinct voice presets, you can match the tone of your brand, whether it is energetic for promo or calm for storytelling. This removes the need for studio bookings.

  • Generate full episodes from text scripts
  • Choose from 8 distinct voice styles
  • Maintain consistent audio quality

E-learning content

Create scalable educational materials without recurring voice talent costs. Use Sage for instructional clarity or Luna for soft storytelling. The API handles up to 4096 characters per request, allowing you to chunk longer lessons efficiently.

  • Ideal for instructional or storytelling tones
  • Handle long texts via character splitting
  • Consistent voice across all modules

Accessibility features

Improve app accessibility by adding spoken text to UI elements. Support 23 languages with auto-detection for Latin scripts. This ensures your product is usable by a global audience without manual localization efforts for audio.

  • Support 23 languages
  • Auto-detect Latin script languages
  • Enhance user accessibility standards

Conclusion

Sound wave illustration for: Conclusion

Speakable provides a reliable, consistent TTS solution tailored for indie developers who need predictable quality without complex configuration. By offering clear voice presets and straightforward integration, we help you avoid the inconsistencies of browser-based APIs while maintaining full control over your audio output.

Our pricing model aligns with startup budgets, charging only for what you use. With no hidden fees or monthly commitments, you can scale your audio features cost-effectively, ensuring that your project’s financial health isn’t compromised by unexpected cloud costs.

Best for consistent quality

We deliver uniform voice characteristics across all generations, ensuring your user experience remains polished and professional. Unlike local browser APIs that vary by device, our cloud engine guarantees the same tone and clarity every time.

Ideal for indie budgets

Start with a free trial of $0.50 credit, then pay just $0.01 per 1,000 characters. There are no subscriptions or minimum spends, making it easy to launch small features or scale up as your user base grows.

Scalable solution

Handle long texts by splitting them into 4096-character chunks, or manage higher traffic with up to 60 requests per minute. Our architecture processes requests sequentially, ensuring stability even during peak usage periods without requiring complex infrastructure management.

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. Node.js documentation
  3. LibriTTS-R paper (arXiv 2305.18802)
  4. LibriVox public-domain audiobooks
  5. openai-node on GitHub
  6. openai-python on GitHub
  7. Creative Commons Attribution 4.0
  8. FFmpeg documentation
  9. curl documentation
  10. MDN: Web Speech API
  11. TRON developer hub: TRC-20
  12. Python documentation

More on this site