Generate Accessible Video Audio with Our Text-to-Speech API

By RK66 Editorial Team · Updated 2026-11-19

This guide helps teams create accessible video content for dyslexic and neurodivergent audiences using the VoxEase TTS API. We explain how to select voices with clear enunciation and neutral tones, adjust speed for readability, and ensure consistent pacing. You will learn to match specific voices like Aria or Atlas to your content type, ensuring the audio supports comprehension without distraction.

We cover practical steps for integration, from choosing between MP3 and WAV workflows to handling multi-language scripts accurately. The guide details uniform pricing, licensing under CC BY 4.0, and technical attributes like pronunciation accuracy. Note that the voices are based on recordings licensed under CC BY 4.0, but the generated audio files themselves do not carry that license restriction. By the end, you will know how to test voice options, avoid common pacing mistakes, and deliver compliant, high-quality audio for your video projects.

Generate Accessible Video Audio with Our Text-to-Speech API

Frequently asked questions

Which voice is the most clear?

Q: Which voice is the clearest? A: Aria is described as a warm, clear narrator voice. While clarity is subjective, Aria is specifically designed for clear narration. You can test any of the eight voices by generating short audio samples via the API to determine which fits your specific content best.

Can I mix male and female voices?

The API processes one voice per request. To mix voices, you must split your text into separate requests, assigning different voices to each part. Then, join the resulting audio files. This allows you to combine any of the eight available voices in your final output.

Do voices sound robotic?

Voices are based on licensed recordings from LibriTTS-R and LibriVox public-domain audiobook readings. They are not synthesized from scratch, offering natural human-like qualities. However, every generated file carries an inaudible watermark marking it as synthetic speech, distinguishing it from original recordings.

Is Aria good for educational content?

Aria is described as a warm, clear narrator voice, making it suitable for educational content. Sage is also described as calm and instructional. You can generate samples with either voice to see which tone best supports your specific educational materials and audience needs.

Is Orion suitable for documentaries?

Yes, Orion is described as a deep, documentary voice. Its tone is specifically suited for documentary-style content. You can generate a sample using the Orion voice to verify its suitability for your specific documentary project before committing to larger volumes of audio generation.

Can I change the speed of Luna?

Yes, you can adjust the speed of any voice, including Luna. The API accepts a speed parameter between 0.25 and 4.0. This allows you to slow down Luna’s soft, storytelling voice for clarity or speed it up for concise content delivery.

Which voice is best for news?

Atlas is described as a neutral, news voice. It is designed for clear, objective delivery suitable for news content. While Atlas is the primary choice, you can test other voices via the API to ensure they meet your specific news format requirements.

Do voices support all 23 languages?

Yes, any of the eight voices can speak any of the 23 supported languages. Supported languages include English, Spanish, German, French, and others listed in the documentation. If the language is not specified, it is guessed from the script, defaulting to English for Latin text.

Is there a difference in price between voices?

No, all voices use the same engine and pricing. The price is $10 per 1 million characters, regardless of which voice you select. You can switch between voices like Aria, Nova, or Felix without any change in cost per character generated.

Can I use Felix for ads?

Yes, Felix is described as an energetic, promo voice. This makes it well-suited for advertising content. You can generate samples using Felix to test how its energetic tone fits your specific ad campaigns and target audience before processing larger volumes of text.

Is Milo too casual for corporate videos?

Milo is described as a friendly, conversational voice. While suitable for casual content, it may work for corporate videos depending on your brand tone. You can generate samples to compare Milo’s conversational style against Atlas’s neutral news tone to determine the best fit.

How do I choose a voice for kids?

You can choose from any of the eight voices. Nova is bright and upbeat, while Luna is soft and storytelling-oriented. Both may suit children’s content. Generate samples with these voices to test which tone best engages your young audience effectively.

Are the voices licensed for commercial use?

Voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. This provides a clear licensing foundation. You can use the generated audio for your projects, including commercial video content, under the terms of these source licenses.

Criteria for Accessible Voice Selection

Sound wave illustration for: Criteria for Accessible Voice Selection

Accessible video content relies on clear audio that reduces cognitive load for dyslexic and neurodivergent viewers. Selecting the right voice requires balancing enunciation clarity with consistent pacing, ensuring each word is distinct without becoming monotonous.

Emotional neutrality prevents distracting affective cues, allowing the message to remain the focus. This approach supports comprehension by minimizing unexpected tonal shifts that can interrupt reading flow.

Clarity and Enunciation

Voices like Aria offer warm, clear narration. This clarity helps listeners track sentence structures more easily, reducing the effort needed to decode speech.

  • Distinct consonant release
  • Clear vowel definition
  • Minimal background noise in audio

Pacing and Speed Control

Adjustable speed (0.25–4.0) lets teams match audio to viewer processing needs. Slower rates aid comprehension, while faster rates suit familiar content. Typical generation speed is roughly real time (about 25 characters per second), though short phrases may take a few seconds and requests are processed sequentially.

  • Fine-tune 0.25x to 4.0x
  • Match cognitive load
  • Avoid rapid-fire delivery

Emotional Neutrality

Voices such as Atlas provide a neutral, news-style delivery that avoids emotional bias. This consistency ensures the content itself drives engagement rather than the performer’s expressive choices.

  • Consistent tone
  • No exaggerated inflection
  • Focus on text

Female Voices Overview

Sound wave illustration for: Female Voices Overview

Selecting the right female voice ensures your video content remains accessible and engaging for neurodivergent audiences. Our three female presets offer distinct tonal qualities, from clear narration to soft storytelling, allowing you to match the emotional tone of your script precisely.

Aria: Warm Narrator

Aria provides a warm, clear narrative voice. Her consistent enunciation reduces cognitive load for listeners with dyslexia, making complex information easier to process without sacrificing warmth.

  • Clear, steady enunciation
  • Warm narrative tone
  • Ideal for long-form educational content

Nova: Bright and Upbeat

Nova brings a bright, energetic quality. Her upbeat delivery keeps viewers engaged and can help maintain attention spans for shorter, dynamic content pieces.

  • Bright, energetic delivery
  • High engagement potential
  • Best for promotional or ad content

Luna: Soft Storyteller

Luna offers a soft, storytelling voice. Her gentle pace is particularly effective for narrative-driven content or calming instructional videos.

  • Soft, gentle tone
  • Narrative-focused delivery
  • Suitable for calming or story content

Male Voices Overview

Sound wave illustration for: Male Voices Overview

Choosing the right male voice depends on your video’s specific tone. Our four male presets offer distinct characteristics suitable for different content types. Each voice is generated from licensed recordings, ensuring consistent quality across all outputs.

The API supports speeds from 0.25x to 4.0x, allowing you to fine-tune pacing for clarity. Whether you need high-energy promotion or steady news delivery, these voices provide a reliable foundation for accessible video content.

Felix: Energetic Promo

Felix delivers an energetic tone ideal for promotional videos. His clear, upbeat delivery keeps viewers engaged during product showcases or quick announcements. This preset works well when you need to convey urgency or excitement without sounding artificial.

Use Felix for short, punchy segments where attention spans are brief. The voice maintains clarity even at higher speeds, making it suitable for fast-paced social media clips or ad copy that requires immediate impact.

  • Best for: Promotional videos, product launches, social media ads
  • Tone: Energetic, upbeat, persuasive
  • Supported languages: All 23 listed languages

Milo: Friendly Conversational

Milo offers a friendly, conversational style that feels natural and approachable. This voice is excellent for explanatory content, tutorials, or community updates where a personal touch enhances understanding. It reduces listener fatigue during longer segments.

Because Milo sounds like a peer rather than a broadcaster, it works well for educational materials aimed at diverse audiences. The natural cadence helps maintain engagement without requiring excessive visual support to retain viewer interest.

  • Best for: Tutorials, community updates, educational explainers
  • Tone: Friendly, approachable, natural
  • Supported languages: All 23 listed languages

Atlas: Neutral News

Atlas provides a neutral, authoritative delivery perfect for news-style content or factual reporting. His consistent tone ensures that information is conveyed without emotional bias, making it ideal for compliance videos or data-driven presentations.

Use Atlas when clarity and neutrality are prioritized over emotional engagement. The voice remains steady across varying speeds, ensuring that complex data or regulatory information is delivered with precision and clarity.

  • Best for: News updates, regulatory compliance, data reports
  • Tone: Neutral, authoritative, clear
  • Supported languages: All 23 listed languages

Specialized Male Voices

Sound wave illustration for: Specialized Male Voices

Male voices provide distinct tonal options for specialized video content, ensuring your audio aligns precisely with the intended audience experience. Selecting the right male preset depends on whether you need conversational engagement, authoritative depth, or clear instructional delivery.

Each voice is engineered for specific communication styles while maintaining the clarity required for accessible materials. Understanding these nuances helps teams avoid mismatched tones that can confuse neurodivergent viewers or dilute the message's impact.

Milo: Conversational Tone

Milo delivers a friendly, conversational tone that feels natural and approachable. This preset is ideal for tutorial videos or explanatory content where a human-like connection helps reduce cognitive load for viewers with dyslexia.

The voice avoids robotic stiffness, making complex instructions easier to follow. Use Milo when you want the audience to feel guided by a supportive peer rather than a distant announcer.

  • Best for: Tutorials, how-to guides
  • Tone: Friendly, conversational

Orion: Deep Documentary

Orion offers a deep, resonant voice suited for documentary-style narratives. Its rich timbre adds weight and seriousness to historical or factual content without sounding overly dramatic.

This preset maintains steady pacing, which aids comprehension for listeners who benefit from consistent auditory cues. It is particularly effective for long-form video essays or educational documentaries.

  • Best for: Documentaries, narratives
  • Tone: Deep, resonant

Atlas: Instructional Clarity

Atlas provides a neutral, news-like delivery that prioritizes clarity and precision. This voice is designed to minimize emotional bias, ensuring that instructional content is delivered with maximum clarity.

Its straightforward articulation makes it an excellent choice for compliance videos, safety instructions, or any content where accurate information transfer is critical. Atlas ensures the message remains the focus, not the voice's personality.

  • Best for: News, instructions
  • Tone: Neutral, clear

Matching Voice to Content Type

Sound wave illustration for: Matching Voice to Content Type

Selecting the right voice ensures your video content meets accessibility standards for neurodivergent viewers. The API lets you match voice characteristics directly to the material type, ensuring clarity and engagement without unnecessary cognitive load.

Each voice preset is designed for specific communication styles. By aligning the voice with the content goal, you maintain consistent tone and improve comprehension across large audiences.

Educational Content

For instructional materials, clarity is paramount. Atlas provides a neutral, clear delivery ideal for news-style explanations. Sage offers a calm, instructional tone.

Use Sage for step-by-step guides or Luna for softer, storytelling-based learning. Both voices support precise enunciation, aiding readers with dyslexia.

  • Use Atlas for neutral, clear explanations.
  • Choose Sage for calm, instructional tone.
  • Luna works well for gentle storytelling.

Promotional Videos

Promotional content benefits from energy and clarity. Felix delivers an energetic, promo-ready voice that captures attention quickly. Nova offers a bright, upbeat tone suitable for modern marketing.

These voices maintain high clarity even at faster speeds, ensuring the message remains understandable without sounding rushed.

  • Felix for energetic, promo-style delivery.
  • Nova for bright, upbeat marketing.
  • Adjust speed to 1.0–1.5 for impact.

Storytelling Narratives

For narrative content, Luna provides a soft, storytelling voice that enhances emotional connection without being overly dramatic. Orion’s deep, documentary tone works well for serious or historical narratives.

Pair these voices with slower speeds (0.75–1.0) to allow listeners time to process complex sentences and visual cues.

  • Luna for soft, engaging stories.
  • Orion for deep, documentary narratives.
  • Use slower speeds for complex text.

Language Considerations

Sound wave illustration for: Language Considerations

Language selection directly impacts clarity for neurodivergent viewers. Our API supports 23 languages, allowing precise control over voice output. Unlike competitors that guess language from script, we let you set the language ISO code explicitly. This prevents mispronunciation in mixed-language texts, ensuring consistent comprehension for dyslexic audiences.

Incorrect language tags can cause voices to default to English, even for Latin-script text. Explicitly setting the language ensures accurate phonetics. Unsupported languages like Vietnamese or Thai are excluded, so verify your content’s language list before integration. This specificity reduces cognitive load for listeners.

Native Speaker Accuracy

Each of our 8 voices speaks all 23 supported languages with native-level accuracy. This includes 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 consistency ensures professional quality across diverse video content.

The system does not guess language if you provide a code. For non-English Latin text, always set the language field. This avoids the default English pronunciation, which can confuse viewers expecting a specific dialect. Unsupported languages are strictly excluded.

  • - Supports 23 specific languages
  • - Explicit language code recommended
  • - No guessing for non-English Latin text
  • - Unsupported: Vietnamese, Thai, Ukrainian

Accent Consistency

Voices maintain consistent accents throughout generation. This stability is crucial for long-form educational videos where listener fatigue is a risk. Inconsistent accents can disrupt the learning process for neurodivergent users. Our API ensures uniform tone and accent per voice preset.

You can adjust speed between 0.25 and 4.0 to match the listener’s processing speed. Slower speeds aid comprehension. Consistent accent paired with controlled pacing enhances accessibility. This combination supports better retention of information in video content.

  • - Stable accent per voice
  • - Speed adjustable 0.25–4.0
  • - Reduces listener fatigue
  • - Supports retention

Multi-Language Scripts

Handle mixed-language videos by setting the language tag per request. Each request handles up to 4096 characters. Split longer texts into parts and join the audio files. This workflow ensures accurate language switching without manual editing. It maintains audio continuity across different language segments.

Use the language parameter to specify the script for each segment. This prevents the default English fallback for Latin text. Ensure all segments use the correct ISO code. This method keeps audio files clean and synchronized with visual cues, enhancing the overall viewing experience.

  • - Max 4096 chars per request
  • - Split and join for long texts
  • - Set ISO code per segment
  • - Avoid default English fallback

Technical Voice Attributes

Sound wave illustration for: Technical Voice Attributes

The API generates synthetic speech using licensed recordings from LibriTTS-R, which influences the naturalness of the output. While the audio is not stored after generation, understanding these technical attributes helps you select the right voice for dyslexic audiences who benefit from clear, consistent enunciation.

Each of the eight voices offers distinct acoustic characteristics. Pitch variation and breath sounds are inherent to the source recordings, providing a human-like quality without requiring complex SSML markup. This simplicity ensures reliable playback across different video editing platforms.

Pitch Variation

Voices like Aria and Luna offer warmer, lower-frequency tones, while Nova and Felix provide brighter, higher-pitched options. This variation allows you to match the emotional tone of your video content without needing external audio processing tools.

The pitch is fixed per voice preset, ensuring consistency across thousands of characters. This stability is crucial for accessibility, as unpredictable pitch shifts can hinder comprehension for neurodivergent viewers.

  • Aria: Warm, clear narrator
  • Nova: Bright, upbeat
  • Luna: Soft, storytelling
  • Felix: Energetic, promo

Breath Sounds

Generated audio includes natural breath pauses between phrases, mimicking human speech patterns. These pauses help segment long sentences, making the content easier to follow for audiences with reading difficulties.

Unlike robotic TTS systems, the licensed recordings provide realistic rhythm. This reduces listener fatigue during longer video segments, supporting better engagement and retention of accessible material.

Pronunciation Accuracy

The system handles 23 languages with accurate pronunciation based on the selected voice. Setting the language parameter ensures correct enunciation for non-English Latin script texts, preventing default English pronunciation errors.

Specialized voices like Atlas (Neutral, News) and Sage (Calm, Instructional) prioritize clarity over expressiveness, making them ideal for instructional content. Their steady rhythm aids comprehension for users who struggle with rapid speech.

  • Supports 23 languages
  • Optional language parameter for accuracy
  • Atlas: Neutral, news style
  • Sage: Calm, instructional style
Technical Voice Attributes

Testing and Iteration

Sound wave illustration for: Testing and Iteration

Testing voice selection requires systematic comparison to ensure maximum accessibility for neurodivergent viewers. By generating identical short clips with different voices, teams can objectively evaluate clarity and fatigue levels before committing to full production.

Iterative refinement relies on direct user feedback and measurable performance data to optimize the final audio output for diverse audiences.

A/B Testing Voices

Generate parallel audio files using the same text but different voices, such as comparing Aria’s warm narration against Luna’s soft storytelling style. Play these side-by-side with dyslexic testers to identify which voice reduces listening fatigue while maintaining clarity.

Use the speed parameter (0.25–4.0) to adjust pacing during this phase, ensuring the voice remains comfortable for extended viewing sessions without rushing or dragging.

  • Compare two voices per script segment.
  • Vary speed settings between 1.0 and 1.5.
  • Record viewer comfort ratings.
  • Select the voice with lowest fatigue scores.

User Feedback Loops

Distribute test clips to a representative sample of your target audience, specifically those with dyslexia or ADHD. Collect qualitative data on pronunciation clarity and emotional neutrality to refine voice choices.

Encourage users to flag specific words or phrases that were difficult to follow, allowing you to adjust text segmentation or voice selection for future batches.

  • Collect subjective comfort scores.
  • Log specific pronunciation issues.
  • Update voice preferences based on feedback.
  • Retest adjusted clips with same users.

Performance Metrics

Track technical metrics such as generation time and character count per request. Since requests are processed sequentially, monitor queue times during high-volume testing to ensure consistent delivery.

Measure the ratio of successful synthesis to total requests. Failed requests are free, allowing you to experiment freely with edge cases or unusual character combinations without financial penalty.

  • Monitor generation queue times.
  • Track success rate of requests.
  • Measure characters per second.
  • Log error frequencies for tuning.

Cost Implications of Voice Choice

Sound wave illustration for: Cost Implications of Voice Choice

Choosing a voice does not affect your billing rate. Every voice preset costs the same flat rate of $10 per 1 million characters, regardless of whether you select the warm Aria or the deep Orion. This uniform pricing simplifies budgeting for large teams producing accessible video content, as you do not need to track costs per voice style.

The high-definition model (tts-1-hd) uses the same engine and incurs no additional charge compared to the standard model. You can upgrade audio quality for clarity without increasing your per-character cost. Volume discounts are not offered; pricing remains consistent regardless of usage volume, making it predictable for steady production workflows.

Uniform Pricing Across Voices

All eight voice presets are priced identically. Whether you generate a script using the bright Nova or the calm Sage, the cost remains $0.01 per 1,000 characters. This consistency allows you to experiment with different tones for dyslexic audiences without worrying about variable costs.

Failed requests are free, ensuring you only pay for successfully generated audio. This protects your budget when iterating on voice selection for specific accessibility needs.

  • - $10 per 1 million characters for all voices
  • - Same rate for tts-1 and tts-1-hd
  • - Failed requests are free

No Extra Cost for HD

The tts-1-hd model provides higher fidelity audio at no extra cost. This is valuable for creating clear, compliant messaging where audio clarity aids comprehension.

Since both models use the same engine, the price difference is zero. You can choose HD for better sound quality without impacting your monthly spend.

Volume Discounts

There are no volume discounts. Pricing is flat at $10 per 1 million characters regardless of how many characters you process.

Balance does not expire, and there are no monthly plans or minimum spends. This structure suits teams with variable production schedules.

  • - No volume discounts
  • - No monthly plans
  • - Balance does not expire

Licensing and Usage Rights

Sound wave illustration for: Licensing and Usage Rights

VoxEase generates audio from licensed recordings, ensuring your video content rests on a clear legal foundation. We handle the infrastructure so you can focus on accessible messaging without worrying about complex rights management for every voice used.

LibriTTS-R License

Our voices are derived from the LibriTTS-R dataset, which uses readings from LibriVox. This means the source material is based on public-domain audiobook readings, providing a robust and verifiable origin for every generated file.

CC BY 4.0 Compliance

The underlying voice data operates under the Creative Commons Attribution 4.0 International license. This standard allows for broad usage while requiring appropriate attribution in your project documentation, ensuring transparency and compliance with open-data standards.

Commercial Use Allowed

You may use the generated audio files for commercial video projects, including paid YouTube channels, corporate training materials, and accessible e-learning platforms. There are no restrictions on selling content created with our API, making it suitable for large team workflows.

  • Use in monetized video content
  • Incorporate into corporate training modules
  • Integrate into accessible e-learning platforms

Integration with Video Editing

Sound wave illustration for: Integration with Video Editing

Integrating VoxEase audio into video workflows is straightforward. Generate speech via our API, then import the file into your editing timeline. This allows precise control over pacing and clarity for accessible content.

Our API returns standard audio formats compatible with major NLEs. Use MP3 for quick previews or WAV for lossless editing quality. The process is simple: request audio, download the file, and place it on your track.

MP3 Workflow

MP3 files are small and load quickly in most editors. Use this format for rapid prototyping or when storage is limited. The audio quality remains clear enough for final export in most web-based video contexts.

  • Supports variable bitrate for balanced file size

WAV Editing

WAV provides uncompressed audio, ensuring no quality loss during multiple edits. This is ideal for precise syncing with visual cues or when applying additional audio effects like volume automation or fades in your NLE.

  • Lossless quality for professional post-production

Syncing with Visuals

Generate short phrases to match specific visual segments. Use the speed parameter (0.25–4.0) to adjust duration if the voice is too fast or slow for your timeline. Trim silence manually after download to align perfectly with video cuts.

Common Mistakes to Avoid

Sound wave illustration for: Common Mistakes to Avoid

Avoiding common pitfalls ensures your accessible videos remain clear and compliant. Small configuration errors can significantly impact comprehension for neurodivergent viewers, so precise settings matter more than creative flair.

Overly Fast Speed

Setting the speed parameter above 2.0 often sacrifices clarity, especially for listeners with processing delays. The API defaults to a natural pace, but pushing it too high makes enunciation difficult to follow.

  • Keep speed between 0.5 and 1.5 for optimal accessibility
  • Avoid maxing out speed for long educational segments

Inappropriate Tone

Using a bright, upbeat voice like Nova for somber news content creates cognitive dissonance. Match the voice preset to the emotional weight of your script to maintain trust and clarity.

Ignoring Language Tags

Latin script text defaults to English pronunciation. If your content is in Dutch, German, or French, you must explicitly set the language ISO code. Otherwise, the voice will guess incorrectly, leading to mispronounced words.

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. This combination supports better retention of information in video content.

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

More on this site