Frequently asked questions
What voices are available?
We offer eight presets: Aria, Nova, Luna, and Sage (female); Felix, Milo, Atlas, and Orion (male). Voices like Aria provide clear narration, while Atlas suits news-style content. All voices are based on licensed recordings and carry an inaudible watermark. You can select any voice for any language supported by the API.
Which voice is best for Arabic?
Any of the eight voices can speak Arabic. For best results, set the language parameter to ar. Without this setting, the system may guess English for Latin-script text. Any voice is suitable for corporate training materials.
Can I use the same voice for all languages?
Yes, every voice can speak all 23 supported languages, including Arabic, Chinese, French, German, Hindi, Japanese, Spanish, and more. Ensure you specify the language ISO code in your request, especially for non-English Latin text, to prevent incorrect language guessing by the system.
What is the speed range?
You can adjust the speed between 0.25 and 4.0. This allows you to slow down complex instructional content or speed up simple announcements. The default speed is 1.0. Adjusting speed does not affect the character count or pricing of your requests.
Is there a voice for kids?
No, we do not offer a specific child voice. Our eight presets are designed for professional, instructional, and documentary styles. If a younger tone is needed, you may adjust the speed or choose a brighter voice like Nova, but no dedicated child voice exists.
What is the character limit?
Each request supports up to 4096 characters. For longer texts, split them into parts and join the resulting audio files. Spaces and punctuation count toward this limit. Failed requests are free, but successful ones consume your balance based on the character count.
Do voices sound robotic?
The voices are based on licensed recordings from LibriVox and LibriTTS-R. They are designed for clear, professional corporate training content. While they carry an inaudible synthetic watermark, they are intended to sound natural for instructional purposes.
Can I mix voices in one file?
No, each request generates audio for a single voice. To mix voices, you must make separate API calls for each voice segment and join the resulting audio files on your end. We do not provide SSML or multi-voice generation in a single request.
What is the HD model?
The tts-1-hd model uses the same engine as tts-1. Both models have the same price of $10 per 1 million characters. Use tts-1-hd when audio clarity is critical for your training content.
Is the API reliable?
We do not provide an SLA or uptime guarantee. Requests are processed sequentially, so high load may cause queuing. Typical speed is roughly real-time. Failed requests are free. We recommend implementing retry logic in your integration for robustness.
How do I handle errors?
If synthesis fails, the charge is returned in full. Check your request parameters, such as valid voice names and supported languages. If issues persist, contact support via email. We do not offer phone support or live chat.
What formats are best for web?
We support MP3, WAV, OPUS, FLAC, and PCM. MP3 and OPUS are efficient for web delivery. WAV and FLAC offer lossless quality. PCM is raw audio. Choose based on your player’s compatibility and bandwidth requirements for your e-learning platform.
Can I use it for videos?
Yes, you can generate audio files and sync them with your video content. The API returns standard audio formats like MP3 or WAV. Since we do not handle video editing, you will need to combine the audio with your visual assets separately.
Is there a free trial?
Yes, signing in with Google gives you $0.50 credit (about 50,000 characters) valid for 7 days. Email sign-ups start without trial credit. You can top up with USDT (TRC-20) for ongoing use. Balances do not expire.
Introduction to Voice Selection
Selecting the appropriate voice for multilingual e-learning directly influences how learners retain instructional material. Clear, professional narration reduces cognitive load, allowing trainees to focus on the content rather than struggling with unclear audio. This clarity is essential for maintaining engagement across diverse global teams.
Impact on Learner Engagement
Consistent voice quality ensures that learners remain focused on the message. A well-chosen voice supports the pacing of information, making complex HR policies or training modules easier to digest. This consistency keeps the audience attentive throughout the module.
Brand Consistency
HR departments often need audio that reflects a professional corporate identity. Using a uniform voice profile across different languages and modules helps establish a recognizable training brand. This uniformity reinforces the credibility of the training materials.
Accessibility Requirements
Clear speech synthesis meets accessibility standards by providing reliable audio alternatives to text. This is crucial for employees with visual impairments or those who prefer auditory learning. It ensures that training content is inclusive and compliant with standard accessibility guidelines.
Voice Catalog Overview
Our voice catalog offers eight distinct presets engineered for clarity and professional tone, ensuring consistent audio quality across multilingual training modules. Each voice is derived from licensed recordings, providing a reliable foundation for corporate e-learning content without the variability of untrained models.
The selection includes four female and four male voices, each optimized for specific instructional styles. This structured approach allows HR agencies to match voice personality to content type, from upbeat promotional training to calm, documentary-style compliance modules.
Female Voices: Aria, Nova, Luna, Sage
Aria provides a warm, clear narration ideal for standard instructional content. Nova offers a bright, upbeat tone suitable for engaging onboarding sequences. Luna delivers a soft, storytelling voice for narrative-driven training. Sage ensures calm, instructional clarity for complex procedural guides.
- Aria: Warm, clear narrator
- Nova: Bright, upbeat
- Luna: Soft, storytelling
- Sage: Calm, instructional
Male Voices: Felix, Milo, Atlas, Orion
Felix brings an energetic, promo-oriented style to dynamic training videos. Milo offers a friendly, conversational tone for interactive modules. Atlas provides neutral, news-style delivery for factual content. Orion delivers deep, documentary-style narration for serious compliance topics.
- Felix: Energetic, promo
- Milo: Friendly, conversational
- Atlas: Neutral, news
- Orion: Deep, documentary
General Characteristics
All voices support 23 languages and can speak any supported language regardless of the voice preset. Generated audio files include an inaudible watermark to identify synthetic speech. Voices are based on licensed LibriTTS-R recordings (CC BY 4.0) and public-domain LibriVox readings, ensuring legal clarity for commercial use.
Voice Profiles for Training
Selecting the right voice for corporate training requires balancing clarity with engagement. Our catalog offers distinct profiles tailored to instructional contexts, ensuring learners retain information effectively across multilingual modules.
The following voices are optimized for training content, providing reliable performance for e-learning, video dubbing, and accessibility text generation.
Sage: Instructional Clarity
Sage delivers a calm, instructional tone ideal for step-by-step guides and procedural training. Her clear articulation reduces cognitive load, making complex instructions easier to follow for diverse audiences.
Use Sage for onboarding materials, software tutorials, and compliance training where precise delivery matters more than emotional flair.
- Best for: Step-by-step instructions
- Tone: Calm and instructional
- Language support: Full 23-language coverage
Atlas: Professional Neutrality
Atlas provides a neutral, news-style voice suitable for general corporate communications. His balanced delivery ensures content remains professional without imposing strong personality traits.
Ideal for policy updates, standard operating procedures, and internal newsletters where consistency and clarity are paramount.
- Best for: Policy updates
- Tone: Neutral and professional
- Use case: Internal communications
Orion: Documentary Depth
Orion offers a deep, documentary-style voice that adds gravitas to historical or contextual training materials. His rich tone engages learners in narrative-driven content.
Recommended for case studies, company history modules, and advanced training where storytelling enhances understanding.
- Best for: Narrative content
- Tone: Deep and documentary
- Use case: Case studies
Language Support and Limitations
The API supports 23 languages, allowing any voice to speak any supported language. This flexibility ensures consistent audio quality across diverse corporate training modules without needing separate voice catalogs for each region.
23 Supported Languages
We support 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, ensuring uniform brand tone across all your multilingual content.
Arabic Pronunciation Quality
All voices handle Arabic natively. Since the engine guesses language from the script, explicitly setting the language parameter is recommended for non-English Latin text to ensure accurate pronunciation and intonation for your learners.
Unsupported Languages List
Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali, and all other languages not listed above are not supported. Ensure your training content stays within the 23 supported languages to avoid delivery issues.
We do not offer custom voice cloning or phoneme dictionaries, so language support is limited to these specific regions.
- Vietnamese
- Indonesian
- Thai
- Ukrainian
- Romanian
- Czech
Technical Configuration
Configuring the TTS API correctly ensures your training audio meets specific delivery standards. By setting the language parameter, you override the default English guess for Latin script text, ensuring accurate pronunciation for non-English modules. This is critical when producing content for diverse international teams.
Adjusting the speed parameter (0.25–4.0) allows you to optimize listening times. For complex instructional content, a slower pace improves comprehension, while faster speeds suit repetitive reviews. Choosing the right audio format balances file size against playback compatibility across your learning management system.
Setting Language Codes
Explicitly define the language field in your JSON request body using ISO codes. Although the API guesses language from script, non-English Latin text defaults to English. Setting the code prevents mispronunciation of proper nouns or mixed-language phrases in HR modules.
This ensures consistent vocal delivery across all regional variants of supported languages, such as Spanish or German, without requiring separate voice selections for each dialect.
- Use ISO 639-1 codes (e.g.,
es,de,fr). - Set
languagefor any Latin-script text outside English. - Omitting the field triggers automatic detection.
Adjusting Speed
The speed parameter accepts values from 0.25 to 4.0. Standard speed (1.0) approximates natural speech. For detailed compliance or technical training, reduce speed to 0.75–0.85 to aid learner retention.
Conversely, increase speed for routine updates or summaries. Note that extreme speeds may affect clarity, so test with your specific content.
Speed adjustments do not alter character billing; you are charged for the input text length regardless of audio duration.
- Range: 0.25 to 4.0.
- Recommended for complex topics: 0.75–0.9.
- Standard narration: 1.0.
Choosing Audio Formats
Select response_format based on your playback environment. MP3 offers universal compatibility with small file sizes, suitable for most LMS platforms. WAV provides lossless quality for professional video dubbing workflows.
Use OPUS for efficient playback in web applications, FLAC for archival storage, or PCM for raw audio processing. Each format has distinct storage and bandwidth implications for your training library.
Ensure your chosen format is supported by the devices your HR staff and trainees will use.
- MP3: Suitable for general LMS use.
- WAV: Lossless quality for video production workflows.
- OPUS: Efficient for web playback.
- FLAC: Archival storage.
- PCM: Raw data processing.
image:
Pricing and Volume Considerations
Pricing is straightforward: $10 per 1 million characters, or $0.01 per 1,000 characters. Spaces and punctuation count toward this total. Failed requests incur no charge, ensuring you only pay for successful synthesis. This model scales linearly with your training volume, making cost prediction reliable for large HR departments.
Cost Per Character
Every character in your input text contributes to the final bill. For typical e-learning scripts, this means long modules may consume significant credits. However, the clear rate allows precise budget forecasting. You can monitor usage directly in your account dashboard to stay within financial limits without hidden fees.
HD vs Standard Model
Both 'tts-1' and 'tts-1-hd' use the same engine and incur the identical price per character. The HD option offers slightly higher audio fidelity. Since the cost is identical, you can switch between models based on listener preference without impacting your budget.
Budget Planning
Start with the $0.50 free trial to estimate character counts for your first course. For ongoing production, top up via USDT (TRC-20) with a $5 minimum. Balance does not expire, allowing you to accumulate credits for seasonal training bursts. With 60 requests per minute, high-volume deployments should batch requests to avoid queue delays.
- Free trial: $0.50 credit (50,000 characters) for 7 days
- Top-up: USDT on TRON network, minimum $5
- No monthly plans or minimum spend requirements
Implementation Workflow
Integrating VoxTrain into your training pipeline begins with configuring the SDK or HTTP client to point at the API base URL. Use your unique API key for authentication and select the tts-1 or tts-1-hd model. Prepare your script text, ensuring it stays within the 4096-character limit per request to avoid truncation issues during playback.
Once configured, send a POST request with your chosen voice preset and preferred audio format. The system returns the audio file directly, allowing you to embed it into your e-learning modules or download it for local storage. This straightforward JSON structure ensures quick setup for development teams.
API Integration Steps
Configure your development environment to use the API base URL. Authenticate using your generated API key. Select the appropriate model (tts-1 or tts-1-hd) and define the voice preset, language, and speed parameters. Ensure your text payload respects the 4096-character limit per request.
Send the POST request to the speech endpoint. The API processes the text and returns the audio file in your specified format (mp3, wav, opus, flac, or pcm). Handle the response by saving the file or piping it directly into your content management system.
- Use the OpenAI SDK with a custom base URL for easy integration.
- Set
Authorization: Bearer your-api-keyin the request headers. - Specify
response_formatto match your LMS requirements (e.g., mp3 for web, wav for high fidelity).
Testing Phase
Before full deployment, run a batch of test scripts across all target languages. Verify that the language code correctly influences pronunciation, especially for non-English Latin script texts. Check audio clarity at various speed settings (0.25–4.0) to ensure comprehension for your specific audience.
Test edge cases such as special characters, numbers, and punctuation. Ensure that the audio generation time aligns with your project timelines, keeping in mind that requests are processed sequentially.
- Verify voice consistency across multiple languages.
- Test speed adjustments for clarity and engagement.
- Check audio format compatibility with your LMS.
Production Deployment
Once testing is complete, integrate the API calls into your production workflow. Monitor usage to stay within the 60 requests per minute limit. If you exceed the rate limit, implement a queuing system to handle requests sequentially without data loss.
Set up alerts for API key usage and balance checks. Since balances do not expire, you can top up with USDT on the TRON network as needed. Ensure your team has access to the API key management page for easy key rotation if necessary.
- Implement request queuing to handle bursts of training content generation.
- Monitor API usage to avoid rate limit issues.
- Set up automated balance checks and top-up alerts.

Quality Assurance
Quality assurance ensures your multilingual training modules meet professional standards before deployment. By systematically validating voice output, character limits, and error states, you guarantee consistent learner engagement across all languages and devices.
Listening Tests
Conduct side-by-side comparisons of selected voices against your brand guidelines. Verify that Sage’s instructional clarity and Atlas’s neutrality remain consistent across different training topics. Listen for natural pacing and ensure the inaudible watermark does not interfere with audio quality.
- Test each voice in all 23 supported languages.
- Verify tone consistency across different training modules.
- Check audio clarity for accessibility compliance.
Character Limit Checks
Ensure each API request contains no more than 4096 characters. Validate that spaces and punctuation are counted correctly toward this limit. Split longer scripts into parts and verify that joined audio files play seamlessly without gaps or overlaps.
- Count characters including spaces and punctuation.
- Test boundary conditions near 4096 characters.
- Verify seamless audio joining for long texts.
Error Handling
Monitor API responses for failed synthesis requests, which are free. Implement retry logic for rate limit errors (60 requests per minute). Ensure your system handles connection timeouts gracefully without duplicating charges or losing audio data.
- Handle 429 Too Many Requests errors.
- Retry failed synthesis requests automatically.
- Log API status codes for debugging.
Data Privacy in Training
For HR training departments, data privacy is critical when processing proprietary course materials. Our API ensures that your input text is used solely for the immediate production of audio and is not retained for model training purposes. This guarantees that your specific training content remains yours, without being used to improve third-party algorithms.
Generated audio files are returned directly in the API response and are not stored on our servers after delivery. This minimizes data exposure and aligns with strict corporate compliance standards, ensuring your training modules remain secure from ingestion to final delivery.
Input Text Usage
Your input text is processed exclusively to generate the requested audio file. It is not used for training our base models, ensuring your proprietary training scripts and HR materials remain confidential.
This distinction is vital for agencies handling sensitive corporate content, as it prevents your specific instructional language from influencing future voice outputs or being used to enhance the base engine.
- Text is not used for model training.
- Content remains proprietary to your organization.
- No third-party algorithmic improvement from your data.
Audio Storage
Generated audio files are streamed directly to your client or server upon completion. We do not store the audio files on our infrastructure after they are delivered.
This ephemeral storage model reduces the risk of data breaches and ensures that your training audio exists only where you choose to save it, whether in your LMS or local archives.
- Audio is not stored on our servers.
- Files are delivered via API response.
- You control final storage location.
Metadata Retention
We retain request metadata, including timestamp, character count, and status, solely for accurate billing and usage tracking. This data is necessary to verify that your account balance reflects your actual consumption.
This metadata does not include the content of your training text, preserving the confidentiality of your instructional material while ensuring transparent and accurate invoicing for your HR department.
- Metadata kept for billing accuracy.
- Includes timestamp, characters, and status.
- Excludes input text content.
Common Use Cases
This section outlines three primary workflows where VoxTrain’s text-to-speech API integrates directly into corporate learning and content pipelines. These use cases leverage our 8 professional voices and 23-language support to streamline production for HR agencies and training departments.
Each scenario below details how to apply specific voice presets and technical configurations to meet distinct content needs, from interactive modules to static accessibility text, ensuring consistent audio quality across multilingual deliverables.
E-Learning Modules
Use Sage for instructional clarity in step-by-step tutorials or Orion for documentary-style background narration. Set the language code explicitly to ensure accurate pronunciation in supported languages like German or Japanese, avoiding default English assumptions.
Split long course scripts into chunks under 4,096 characters to maintain consistent audio quality. The 0.25–4.0 speed range allows you to adjust pacing for complex technical terms, ensuring learners can follow along comfortably.
- Use Sage for clear, instructional narration.
- Apply Orion for documentary-style segments.
- Explicitly set ISO language codes.
- Split text to 4,096 characters max.
Video Dubbing
Generate audio files in MP3 or WAV format to sync seamlessly with existing video tracks. With support for 23 languages, you can produce localized versions of training videos without hiring multiple voice actors.
Maintain brand consistency by using the same voice preset across all language versions. For example, use Atlas for neutral news-style updates or Felix for energetic promotional training clips, ensuring a unified corporate voice.
- Output MP3 or WAV formats.
- Support for 23 languages.
- Maintain consistent voice presets.
- No need for multiple voice actors.
Accessibility Text
Convert text-based training materials into audio for visually impaired employees. Use Aria or Nova for clear, engaging reading of manuals or policies. The API handles up to 4,096 characters per request, simplifying batch processing.
Ensure compliance with accessibility standards by providing high-quality synthetic speech. The inaudible watermark confirms synthetic origin, which may be required for transparency in corporate communications. No streaming is needed; download the full audio file.
- Convert manuals to audio.
- Use Aria or Nova presets.
- Handle 4,096 chars per request.
- Inaudible watermark confirms synthetic origin.
Troubleshooting
Production deployments often encounter predictable hurdles. This guide addresses the three most frequent issues: rate limits, language detection errors, and audio format mismatches.
Resolve these quickly to maintain smooth e-learning content pipelines without interrupting learner access.
Rate Limit Issues
Your key allows 60 requests per minute. If you exceed this, requests queue or fail. Failed requests are free, but successful ones consume your balance.
Monitor your dashboard for usage spikes. Split large texts into parts under 4096 characters to stay within efficient batch sizes.
- Limit: 60 requests/minute per key
- Pre-top-up limit: 10 requests/minute
- Failed requests incur no charge
Language Detection Errors
If you omit the language field for non-English Latin text, the system defaults to English, causing pronunciation errors.
Always set the language parameter to the ISO code for non-English content to ensure accurate synthesis.
Audio Format Issues
Select mp3, wav, opus, flac, or pcm based on your LMS requirements. Ensure your player supports the chosen format.
If audio fails to play, verify the response_format matches your player’s capabilities and check for valid JSON responses.
Conclusion and Next Steps
Selecting the right voice and configuring the API correctly ensures your multilingual training modules are clear, consistent, and accessible. By leveraging the eight professional presets and adhering to the 4096-character limit, you can maintain high audio quality across diverse linguistic contexts without unnecessary complexity.
With precise language codes and speed adjustments, your HR teams can produce scalable e-learning content efficiently. The straightforward pricing model and reliable API integration allow you to focus on content delivery rather than technical infrastructure, ensuring a seamless learner experience.
Summary of Best Practices
Stick to the recommended voices like Sage for instructional clarity or Atlas for neutral professionalism. Always specify the language code to avoid detection errors, and split long texts into 4096-character chunks to ensure consistent generation speed and quality.
- Use Sage or Atlas for standard corporate training.
- Set explicit language codes for non-English text.
- Split input text at 4096 characters.
Getting Started
Visit the /get-api-key/ page to create your account via Google or email. Generate your API key and use the provided curl example to make your first request. Start with the free trial credit to test voice selection and audio quality before scaling your production volume.
Support Contact
For technical questions or billing inquiries, reach out via email. We do not offer phone support or live chat; our team responds to all emails promptly to assist with API integration or voice selection issues.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Ensure your chosen format is supported by the devices your HR staff and trainees will use.
23 languages
Any voice speaks any of these languages.
