How to Use Our TTS API for Audiobook Production?

By ReadCast editorial team · Updated 2026-11-19

This guide details how independent authors and audiobook narrators can use the ReadCast API to convert manuscripts into professional audio files. We cover the complete workflow, from preparing clean text and selecting the appropriate voice preset to handling batch requests and managing rate limits. You will learn practical strategies for segmenting long documents to fit character constraints and optimize generation speed.

Read this to understand the specific benefits and limitations of character-based TTS for self-publishing. We provide concrete examples for fiction narration and educational content, explain how to estimate production costs accurately, and clarify licensing details for the underlying voice data. Whether you are producing a single title or a full series, this page gives you the technical knowledge to integrate reliable, consistent audio generation into your publishing pipeline.

How to Use Our TTS API for Audiobook Production?

Frequently asked questions

Is TTS suitable for fiction audiobooks?

Yes, voices like Luna offer soft, storytelling tones ideal for fiction. The API supports 23 languages and allows speed adjustments from 0.25 to 4.0. While not perfect, the synthetic speech is suitable for consistent narration. You can generate chapters individually, ensuring quality control for your manuscript without needing human narrators for every project.

How do I handle proper nouns in my text?

The API does not support SSML or custom pronunciation lexicons. You must ensure correct spelling in your input text, as the system guesses language based on script. For non-English Latin text, explicitly set the language parameter. Proper nouns are read as written, so accurate input is essential for correct pronunciation.

Can I change the speed of the narration?

Yes, you can adjust the speed between 0.25 and 4.0 using the speed parameter. This allows you to fine-tune the pacing for your audiobook. Faster speeds reduce generation time, while slower speeds may improve clarity. The API processes requests sequentially, so speed adjustments do not affect the queue order.

Do I need to edit the audio manually?

The API returns complete audio files, but lacks SSML for precise control. You may need to split long texts into parts under 4096 characters and join them. Since requests are processed sequentially, you might need to manage timing. Manual editing ensures seamless transitions between chapters or sections generated in separate requests.

How long does it take to generate a 10-hour audiobook?

Generation takes roughly real-time, about 25 characters per second. A 10-hour audiobook requires approximately 10 hours of generation time. Requests are processed sequentially, so under load, your request may wait in a queue. Plan for potential delays if multiple users are generating audio simultaneously.

Can I use multiple voices in one book?

Yes, you can switch voices between chapters or sections by specifying different voice parameters in each request. You can use eight presets like Aria, Nova, or Felix. Each request generates a separate audio file, which you can combine manually. This allows for distinct character voices or narrator changes.

What is the best audio format for distribution?

Which audio format is best for distribution? The API supports mp3, wav, opus, flac, and pcm. MP3 offers wide compatibility, while WAV and FLAC provide lossless quality. Opus ensures efficient compression. Choose based on your platform requirements and storage needs. All formats are generated directly from the API response, ensuring consistent quality across your audiobook files.

Do I need a website to sell my audiobooks?

No, the API generates audio files you can distribute anywhere. You can sell on platforms like Audible, iTunes, or your own site. The API does not handle sales or distribution. You manage your own distribution pipeline. The API simply provides the audio file, which you can then upload to any retailer.

How do I handle pauses and breaks in text?

The API does not support SSML tags for precise pause control. You can use punctuation in your text to influence natural breaks. For longer pauses, you may generate separate audio segments and join them manually. Adjusting the speed parameter can also subtly affect the pacing of the narration.

Is the voice robotic or natural sounding?

Voices are based on licensed recordings from LibriTTS-R and LibriVox. They sound natural but carry an inaudible watermark marking them as synthetic. They are not perfect human replicas but offer high quality. The tone is clear and consistent, suitable for audiobooks, though some listeners may detect the synthetic nature.

Can I regenerate a specific chapter if I'm unhappy?

Yes, you can regenerate any chapter by sending a new request with the same text and parameters. Failed requests are free, but successful regenerations count against your character limit. You can issue a new API key if needed. Each request is independent, allowing you to retry until satisfied with the output.

What if the language detection is wrong?

If the language is not set, the API guesses based on the script. Latin text defaults to English. For non-English Latin text, explicitly set the language parameter. This ensures correct pronunciation. If detection fails, manually specifying the ISO code resolves the issue and improves accuracy for your specific language.

Do I get a certificate for the audio quality?

No. ReadCast API does not provide certificates for audio quality or formal quality assurance documentation. We return standard audio files (MP3, WAV, OPUS, FLAC, PCM) generated from your text and voice settings. While we ensure consistent synthesis based on your input, we do not offer independent certifications for the output audio, so you manage quality control yourself.

Can I use the audio in YouTube videos?

Yes, you can use the generated audio in YouTube videos. The API returns audio files that you can embed or upload. Note that the audio contains an inaudible watermark marking it as synthetic. Ensure you comply with YouTube's policies regarding synthetic media. The API does not handle rights or distribution for your content.

Understanding TTS for Audiobooks

Sound wave illustration for: Understanding TTS for Audiobooks

Text-to-Speech (TTS) converts written text into audio files through a cloud API. For self-publishers, this technology removes the barrier of hiring voice actors or recording studios, allowing you to produce audiobooks directly from your manuscript files. The process is straightforward: send your text to the API, and receive a standardized audio file ready for distribution.

This approach offers significant benefits for independent authors by lowering upfront costs and accelerating production timelines. You retain full control over the narrative voice and can adjust speed or format without negotiating contracts. It is a practical solution for getting your book into listeners’ ears quickly, using a reliable, character-based pricing model that scales with your output.

What is Text-to-Speech?

TTS is a cloud service where you send text and receive an audio file. It uses neural networks to synthesize speech from written input. You do not record sound; you generate it programmatically. This method ensures consistent audio quality across all chapters, eliminating variations in human performance.

  • Cloud-based processing
  • Programmatic generation
  • Consistent output quality

Benefits for Self-Publishers

Self-publishers gain full creative control over their audiobooks. You can choose from eight distinct voices, each suited to different genres. The pricing is transparent: $0.01 per 1,000 characters. There are no monthly fees or minimum spends, making it ideal for authors with irregular release schedules.

  • Low upfront cost
  • No monthly fees
  • Full creative control
  • Scalable pricing

Limitations to Know

The API processes requests sequentially, so long texts may take time. Each request is limited to 4,096 characters, requiring you to split longer manuscripts. There is no SSML support for fine-tuning pauses or emphasis. You must handle text segmentation and audio joining in your own workflow.

  • Sequential processing
  • 4,000 character limit per request
  • No SSML support

Choosing the Right Voice

Sound wave illustration for: Choosing the Right Voice

Selecting the appropriate voice is critical for maintaining listener engagement throughout a full-length audiobook. The eight available presets span distinct tonal ranges, from the warm clarity of Aria to the neutral reliability of Atlas. Choosing incorrectly can cause listener fatigue, especially during multi-hour sessions.

The API allows you to switch voices per chapter if your narrative structure demands it, though consistency is generally preferred. Each voice is trained on specific licensed recordings, ensuring a consistent timbre across your entire production.

Matching Genre to Genre to Voice

Match the voice character to your book's genre. For historical fiction, Orion provides a deep, documentary-style narration suitable for serious themes. Contemporary romance or light fiction may benefit from Nova’s bright, upbeat delivery or Luna’s soft storytelling approach.

Non-fiction works often require Sage’s calm, instructional tone or Felix’s energetic promo style to keep the audience attentive. Mystery and thriller narratives might find Atlas’s neutral news anchor style helps maintain tension without distracting emotional inflections.

  • - Aria: Warm, clear narrator ideal for literary fiction
  • - Nova: Bright, upbeat for contemporary fiction or self-help
  • - Luna: Soft, storytelling voice for fantasy or romance
  • - Sage: Calm, instructional tone for non-fiction
  • - Felix: Energetic, promo style for marketing or short reads
  • - Milo: Friendly, conversational for personal narratives
  • - Atlas: Neutral, news-style for factual content
  • - Orion: Deep, documentary voice for history or serious topics

Audience Preferences

Consider your target demographic. Younger audiences may prefer the brighter tones of Nova or Milo, while older listeners often appreciate the steady, clear articulation of Aria or Atlas.

Any voice can speak any of the 23 supported languages, allowing you to adapt the same story for different markets. If your manuscript is in English but targets non-native speakers, ensure the selected voice’s clarity matches the expected proficiency level.

Testing Samples

Generate short 4096-character samples using the /v1/audio/speech endpoint to test how different voices handle your specific text. Pay attention to pacing, especially with complex proper nouns or punctuation.

Since requests are processed sequentially, test a few voices during low-load periods to gauge natural speed. Adjust the speed parameter between 0.25 and 4.0 to fine-tune the listening experience before committing to full chapter production.

Preparing Your Manuscript

Sound wave illustration for: Preparing Your Manuscript

Preparing your manuscript for text-to-speech requires careful attention to detail. Self-published authors must ensure their text is clean and properly formatted before sending it to the API. This step ensures consistent audio quality and prevents unexpected pauses or errors during generation.

Text Cleaning

Remove unnecessary line breaks and ensure consistent spacing. The API processes text up to 4096 characters per request, so clean manuscripts reduce preprocessing effort.

Handling Special Characters

Standard punctuation works well, but complex symbols may need simplification. The API does not support SSML, so avoid custom pronunciation tags. Test special characters in short samples before batch processing.

  • Replace ellipses (...) with standard periods or commas
  • Convert symbols like © or ® to words if they disrupt flow
  • Ensure consistent capitalization for proper voice pacing

Segmentation Strategy

Split long texts into segments of up to 4096 characters. This ensures each request is processed efficiently and avoids hitting rate limits too quickly. Short phrases may take a few seconds to generate, so plan your workflow accordingly.

  • Divide chapters into manageable chunks
  • Keep paragraphs intact to maintain natural speech flow
  • Use the API’s language detection or specify ISO codes for non-English texts

Generating Audio Files

Sound wave illustration for: Generating Audio Files

Generating audio from your manuscript requires sending a POST request to the speech endpoint with your JSON payload. You must provide a valid model (tts-1 or tts-1-hd), the input text, and a voice preset. The API returns the audio file directly in the response, ready for download or immediate storage in your production pipeline.

For long manuscripts, you split the text into chunks of up to 4096 characters. Each chunk generates a separate audio file that you later join. This ensures stable processing and avoids truncation errors during synthesis.

API Request Structure

Send a POST request to https://api.link02.info/v1/audio/speech with a JSON body. Include the model ("tts-1" or "tts-1-hd"), input text, voice preset, and response_format (mp3, wav, opus, flac, or pcm). Optionally set speed between 0.25 and 4.0 and specify a language ISO code if the script is ambiguous.

  • Use tts-1 for standard quality or tts-1-hd for higher fidelity.
  • Include Authorization: Bearer your-api-key in the headers.
  • Text input is limited to 4096 characters per request.

Handling Rate Limits

Your API key allows 60 requests per minute once funded, or 10 requests per minute before your first top-up. If you exceed this limit, the API will return an error. Plan your batch sizes accordingly to avoid delays in generating long chapters.

  • Standard limit: 60 requests/minute.
  • Pre-top-up limit: 10 requests/minute.
  • Failed requests do not count toward your character usage.

Batch Processing

Split your manuscript into segments of 4096 characters or less. Send each segment as a separate request. Once you receive all audio files, concatenate them in the correct order to create the final chapter audio. This method handles large texts reliably without losing context.

  • Split text by natural breaks (paragraphs, sentences) under 4096 chars.
  • Process requests sequentially or in parallel within rate limits.
  • Join output files using any standard audio editor.

Post-Processing and Quality Control

Sound wave illustration for: Post-Processing and Quality Control

After synthesis, select the output format that balances file size and quality. MP3 offers efficient compression for distribution, while WAV or FLAC preserves higher fidelity for archival purposes. The API returns the audio file directly in the response, ready for immediate download.

Generated audio is not stored by the API, so you must handle local storage and editing. Minor manual adjustments may be needed to fix timing gaps between chapters or normalize volume levels across different voices. The audio carries an inaudible watermark to identify it as synthetic speech.

Audio Format Selection

Choose MP3 for smaller file sizes suitable for distribution, or WAV/FLAC for higher fidelity masters. PCM is available for uncompressed raw data. Format choice does not affect character pricing.

  • MP3: Suitable for web distribution and smaller storage footprints.
  • WAV/FLAC: Preferred for high-quality archival masters.
  • OPUS: Efficient compression for voice-heavy applications.

Manual Editing Needs

The API returns discrete audio files per request. You must concatenate chapters manually or via script. Speed adjustments between 0.25x and 4.0x allow fine-tuning narration pace, but extreme speeds may reduce clarity.

Since the audio is returned directly and not stored, you are responsible for saving files. Minor editing may be required to smooth transitions between segments or adjust volume consistency.

  • Concatenate audio segments for full chapters.
  • Adjust speed (0.25–4.0) for pacing control.
  • Normalize audio levels in post-production.

Watermark Identification

Every generated file contains an inaudible watermark marking it as synthetic speech. This helps identify AI-generated content in distribution channels. The watermark is embedded in the audio data and does not affect playback quality.

  • Inaudible synthetic speech marker.
  • Embedded in all output formats.
  • Used for content identification.

Pricing Your Production

Sound wave illustration for: Pricing Your Production

Accurate character counting is the foundation of predictable audiobook production costs. Since our API charges $0.01 per 1,000 characters—including spaces and punctuation—you must account for the entire manuscript length. This precise metric ensures you know exactly how much each file will cost before generating it.

Estimating Character Counts

Use standard word processors or code editors to count characters in your text files. Remember that every space, comma, and paragraph break contributes to the total. For a standard 50,000-word novel, expect approximately 250,000–300,000 characters, resulting in a base cost of $2.50–$3.00.

  • Include all punctuation and spaces in your count
  • Test a 1,000-character sample to verify accuracy
  • Account for special characters or symbols in your manuscript

Cost per Chapter

Calculate costs per chapter to manage cash flow and pricing strategies. If a chapter contains 5,000 characters, it costs $0.05 to generate. This low per-unit cost allows you to experiment with different voices or speeds without significant financial risk, enabling iterative refinement of your audiobook.

  • Break down large manuscripts into chapter-sized batches
  • Use failed requests to refine voice settings without penalty
  • Track character usage per voice for consistent budgeting

Budgeting for Series

For series, multiply your per-book cost by the number of volumes. With a $0.50 free trial, you can test your workflow immediately. For a 10-book series averaging $3.00 each, total production costs remain under $30, making self-publishing highly accessible without upfront enterprise contracts or subscriptions.

  • Scale costs linearly with the number of books
  • Leverage the free trial for initial voice testing
  • Monitor usage via the dashboard to avoid surprises

Legal and Licensing Aspects

Sound wave illustration for: Legal and Licensing Aspects

ReadCast API leverages voice data from LibriTTS-R (CC BY 4.0) and public-domain LibriVox audiobooks, providing a clear legal foundation for content creation. Understanding these source licenses ensures that your generated audio files are safe for commercial distribution without unexpected copyright conflicts.

Every generated file carries an inaudible watermark that marks it as synthetic speech. This separation simplifies rights management for self-publishers, allowing you to focus on distribution rather than navigating complex voice-acting union agreements or individual artist contracts.

Source Voice Licenses

The underlying voice models are trained on LibriTTS-R, licensed under Creative Commons BY 4.0. Original recordings come from LibriVox, which consists of public-domain audiobook readings. This combination means the synthetic output inherits a permissive license structure suitable for commercial use.

  • LibriTTS-R: CC BY 4.0
  • LibriVox: Public Domain
  • No additional clearance needed for standard usage

Commercial Usage Rights

Generated audio files can be used for commercial audiobooks, podcasts, and educational materials without extra fees. The API does not impose restrictions on the number of copies or platforms where your audio is published. You retain full rights to the final audio file you produce.

  • Unlimited commercial distribution
  • No royalty payments per download
  • Suitable for Amazon KDP, Audible, and Spotify

Attribution Requirements

While the CC BY 4.0 license technically requires attribution, ReadCast API does not mandate specific credit lines for end-user audiobooks. However, it is good practice to note that the audio was generated using AI voices sourced from LibriVox. This transparency helps maintain trust with your listeners.

  • Attribution is not mandated by ReadCast
  • Recommended: 'AI-generated from LibriVox voices'
  • Clear labeling enhances listener trust
Legal and Licensing Aspects

Workflow Optimization

Sound wave illustration for: Workflow Optimization

Efficient audiobook production relies on automating the repetitive tasks of text segmentation and API calls. By scripting the 4096-character split and queue management, you ensure consistent delivery without manual intervention for every chapter.

Leveraging the OpenAI SDK simplifies integration by requiring only a base URL change. This allows you to use familiar code structures while benefiting from our specific voice presets and pricing model.

Automating with Scripts

Write a simple script to divide your manuscript into chunks under 4096 characters. Send each chunk via POST requests, then concatenate the returned audio files. This handles long documents seamlessly without hitting input limits.

  • Split text by character count, not sentences, for predictable audio length.
  • Concatenate outputs sequentially to maintain narrative flow.
  • Handle failures by retrying specific segments without restarting the whole book.

Using the OpenAI SDK

Configure your SDK client with base_url set to https://api.link02.info/v1. Use the standard audio/speech endpoint with your ReadCast API key. This ensures compatibility with existing workflows while accessing our 8 licensed voices.

Monitoring Usage

Track your character consumption and balance directly on the /get-api-key/ dashboard. Since requests are processed sequentially, monitor queue times during peak usage to adjust your publishing schedule accordingly.

Common Challenges and Solutions

Sound wave illustration for: Common Challenges and Solutions

Audiobook production often involves navigating technical constraints that affect workflow efficiency. Understanding these common hurdles allows independent authors to maintain consistent quality without unexpected delays or errors in their final audio files.

Handling Long Documents

The API accepts up to 4096 characters per request. For longer manuscripts, you must split the text into segments and join the resulting audio files. This approach ensures stability and prevents timeouts during synthesis.

  • Split text into chunks under 4096 characters.
  • Concatenate outputs sequentially for a complete chapter.

Language Detection Issues

If the language field is omitted, the system guesses based on script, defaulting to English for Latin text. For non-English Latin-script content, always set the language parameter explicitly to avoid mispronunciation.

  • Set language explicitly for Latin-script non-English text.
  • Supported: 23 languages including Arabic, Chinese, and Spanish.

Speed Adjustments

Adjustable speed ranges from 0.25x to 4.0x. Use lower speeds for complex educational content or higher speeds for fast-paced fiction. Note that requests are processed sequentially, so load may cause queuing delays.

Tools and Integrations

Sound wave illustration for: Tools and Integrations

ReadCast API integrates seamlessly into existing content creation workflows. Authors can export clean text from standard editors, ensuring metadata and formatting remain intact before API submission.

Text Editors

Most word processors and Markdown editors work well. Simply copy the cleaned manuscript text and paste it into your API client or script.

  • Plain text (.txt) or Markdown (.md) formats are recommended.
  • Ensure special characters are encoded correctly before sending.

Audio Players

Standard media players support all output formats (mp3, wav, opus, flac, pcm). Use any player that handles standard audio codecs for quality checks.

Distribution Platforms

Generated audio files are ready for direct upload to ACX, Spotify, or Apple Books. No conversion is needed for standard formats like MP3 or WAV.

Case Studies and Examples

Sound wave illustration for: Case Studies and Examples

Real-world application of the ReadCast API demonstrates its utility across distinct content genres. By leveraging specific voice presets and character limits, creators can maintain narrative consistency while adhering to the 4096-character constraint per request.

Fiction Narration

Using 'aria' or 'nova' for light fiction allows for clear, engaging narration. Writers split manuscripts into segments under 4096 characters, ensuring each API call completes efficiently. The resulting audio files maintain a consistent tone, suitable for self-publishing platforms.

  • Voice: aria (warm narrator) or nova (upbeat)

Non-Fiction Clarity

For instructional content, 'sage' or 'atlas' provide neutral, clear delivery. This approach suits technical guides or business books where clarity outweighs emotional nuance. The API handles standard punctuation well, reducing post-production cleanup time.

  • Voice: sage (calm, instructional) or atlas (neutral, news)

Educational Content

Educational materials benefit from the 'milo' conversational style. Teachers can generate audio for lessons quickly, using the 0.25–4.0 speed range to adjust pacing for different learning levels. Each file includes a synthetic watermark for identification.

  • Voice: milo (friendly, conversational)

Future-Proofing Your Content

Sound wave illustration for: Future-Proofing Your Content

Build your audiobook pipeline with confidence. The API handles volume efficiently, allowing you to process entire series without infrastructure limits. Character-based pricing scales linearly, so your costs remain predictable as your catalog grows from one title to hundreds.

Scalability

The system processes requests sequentially, ensuring stable output even during high-volume generation. With 60 requests per minute, you can automate batch jobs for long manuscripts without manual intervention.

  • Linear cost structure supports large catalogs
  • Automated batching for multi-chapter books

Voice Consistency

Maintain a uniform narrator voice across your entire series. Since you select specific voices like Aria or Atlas, every chapter retains the same tonal quality, essential for listener immersion in serialized fiction.

  • Same voice across all chapters
  • No drift between sessions

Technology Updates

Your content remains compatible with standard audio formats like MP3 and WAV. As standards evolve, these widely supported formats ensure your audiobooks play on any modern device or distribution platform without re-encoding.

  • Standard formats: MP3, WAV, OPUS
  • No proprietary lock-in

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. The ability to specify language codes ensures accurate pronunciation for non-native speakers, enhancing accessibility for diverse student demographics.

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. LibriTTS-R corpus (OpenSLR 141)
  3. OpenAI API reference: create speech
  4. Creative Commons Attribution 4.0
  5. Tether (USDT)
  6. LibriTTS-R paper (arXiv 2305.18802)
  7. LibriVox public-domain audiobooks
  8. openai-node on GitHub
  9. Python documentation
  10. curl documentation
  11. IETF RFC 7845: Ogg Opus
  12. ISO 639 language codes

More on this site