Frequently asked questions
What is the best voice for training?
Q: Which voice is best for training? A: Sage offers a calm, instructional tone ideal for corporate training. Atlas provides a neutral, news-style delivery suitable for clear information transfer. Both are effective for e-learning content, allowing you to choose based on whether you prefer a conversational instructional style or a formal news presentation.
Which voice is best for Arabic?
All eight voices support Arabic. You can select any preset, such as Sage for instructional content or Atlas for news-style delivery. Ensure you set the language parameter to 'ar' in your request, as the system defaults to English for Latin-script text.
What is the difference between Sage and Atlas?
Sage is designed for calm, instructional content, making it suitable for step-by-step training modules. Atlas offers a neutral, news-style tone, which is better for formal announcements or factual reporting. Choose Sage for educational guidance and Atlas for objective information delivery.
Can I use Orion for news?
Orion is described as deep and documentary-style, making it suitable for narrative news or deep-dive content. For standard news delivery, Atlas provides a more neutral tone. Orion works well if you want a more immersive, documentary feel for your corporate materials.
Does voice affect price?
No. All voices use the same engine models: tts-1 or tts-1-hd. Pricing is based solely on character count, at $0.01 per 1,000 characters. You can switch between any of the eight voices without any change in cost or billing structure.
Can I change speed?
Yes. You can adjust the speed parameter between 0.25 and 4.0. This allows you to slow down complex training instructions for better comprehension or speed up repetitive content. The API accepts this parameter in the JSON body of your request.
Is Aria good for narration?
Aria is described as warm and clear, making her well-suited for narration. She is ideal for e-learning content that requires a friendly yet professional tone. Her clear delivery helps maintain listener engagement during longer training modules or instructional videos.
What is Nova's tone?
Nova has a bright and upbeat tone. She is best used for energetic content, such as product introductions or engaging training segments. Her style is less formal than Sage or Atlas, making her suitable for lighter, more dynamic corporate communications.
Is Luna suitable for stories?
Yes. Luna is designed for soft, storytelling content. She is ideal for narrative-driven training modules or case studies where a gentle, engaging tone enhances the listener's experience. Her style is distinct from the more instructional Sage or news-like Atlas.
Can Felix be used for promos?
Yes. Felix is energetic and promo-focused. He is well-suited for promotional training modules, announcements, or content requiring high energy. His style is distinct from the calm instructional tone of Sage, making him ideal for dynamic corporate communications.
What is Milo's style?
Milo is friendly and conversational. He is ideal for interactive training modules or customer service simulations. His style bridges the gap between formal news delivery and casual narration, making him suitable for approachable corporate training content.
Do voices support all languages?
Each voice can speak any of the 23 supported languages, including Arabic, English, and German. You do not need a specific voice for a language; you simply set the language parameter. All voices support the same language set.
Is there a demo available?
The API returns audio files directly. You can test voices by sending a POST request to /v1/audio/speech with the desired voice parameter. There is no built-in web player; you must generate audio to hear the specific tone and speed settings.
How do I test voices?
Send a POST request to the /v1/audio/speech endpoint with your text and the desired voice ID. You can view available voices and their details via the GET /v1/voices endpoint. Use the provided SDK or curl to generate and download the audio file.
Understanding Voice Personalities
Choosing the right voice ensures corporate training materials are clear and engaging. For Arabic content, three presets stand out for their professional clarity and instructional tone. Sage, Atlas, and Orion each serve distinct purposes in e-learning environments.
These voices are based on licensed recordings and support 23 languages, including Arabic. They avoid emotional exaggeration, focusing instead on clarity for instructional and documentary-style content.
Sage: Calm and Instructional
Sage offers a calm, instructional tone ideal for soft skills training. Its clear narration helps learners absorb procedural steps without distraction.
This voice works well for compliance modules where steady pacing aids retention.
- Best for: step-by-step instructions
- Tone: Calm, instructional
- Language: Arabic supported
Atlas: Neutral and News-Like
Atlas provides a neutral, news-like delivery suitable for technical documentation. Its consistent tone ensures factual content remains objective.
Ideal for corporate updates or policy explanations where bias must be minimized.
- Best for: Technical docs
- Tone: Neutral, news-like
- Language: Arabic supported
Orion: Deep and Documentary
Orion’s deep, documentary style suits longer training courses. Its authoritative presence keeps learners engaged during extended content.
Use for case studies or historical training modules requiring a serious tone.
- Best for: Documentaries
- Tone: Deep, authoritative
- Language: Arabic supported
Matching Voice to Content Type
Choosing the right voice depends on the cognitive load of your training material. For soft skills, a warmer tone helps build rapport, while technical content requires neutrality to prevent distraction. Alignment between voice personality and content type ensures better retention and clearer instruction delivery.
Soft Skills Training
Use Aria or Luna for interpersonal modules. Their warm, clear, or soft tones make feedback and scenario-based learning feel approachable.
- Aria for clear narration
- Luna for storytelling
Technical Documentation
Select Atlas or Felix for procedural content. Atlas provides a neutral, news-like delivery ideal for step-by-step guides. Felix adds energy to promotional or quick-reference technical overviews.
- Atlas for neutrality
- Felix for energy
Compliance Modules
Orion’s deep, documentary style suits serious regulatory content. It conveys authority and gravity without sounding aggressive. Use Sage for instructional compliance updates requiring calm, steady pacing.
- Orion for authority
- Sage for calm instruction
Arabic Language Specifics
Selecting the right voice for Arabic e-learning requires balancing clarity with instructional tone. Our catalog includes voices that handle Arabic script effectively, ensuring consistent delivery across training modules.
Voice Suitability for Arabic
All eight preset voices support Arabic. For instructional content, Sage offers a calm, instructional tone ideal for soft skills. Atlas provides a neutral, news-like delivery suitable for compliance modules. Orion’s deep, documentary style works well for complex technical explanations.
- Sage: Calm and instructional
- Atlas: Neutral and news-like
- Orion: Deep and documentary
Pronunciation Clarity
Voices are based on licensed recordings, providing clear articulation. The API guesses the language from the script, but explicitly setting the language field ensures accurate Arabic pronunciation. This prevents defaults to English phonetics for Latin-script inputs.
Cultural Neutrality
Our voices offer a neutral tone suitable for multinational corporations. This ensures the content feels professional and universally accessible to Arabic speakers across different regions, without favoring a specific dialect.
Technical Constraints
Technical constraints define how your Arabic training audio is generated and delivered. Understanding speed limits, format choices, and character boundaries ensures consistent quality across your e-learning modules without unexpected errors or delays during production.
Each constraint impacts the final output’s clarity and file size, which matters when distributing content to diverse devices or bandwidth-limited learners in corporate environments.
Speed Adjustment Range
You can adjust playback speed between 0.25 and 4.0 times normal rate. For Arabic instructional content, speeds between 0.8 and 1.5x maintain clarity without distorting pronunciation. Faster rates may reduce listening comprehension for non-native speakers.
Testing different speeds helps balance production efficiency with learner retention. Slower speeds improve understanding for complex technical terms, while faster speeds suit repetitive compliance reviews.
- Minimum speed: 0.25x
- Maximum speed: 4.0x
- Recommended range for Arabic: 0.8–1.5x
Audio Format Impact
Choose from mp3, wav, opus, flac, or pcm based on your distribution needs. MP3 offers small file sizes for web delivery, while WAV and FLAC preserve higher fidelity for archival or high-quality playback systems.
Opus provides efficient compression, whereas PCM is uncompressed raw audio. Select formats that align with your learning management system’s capabilities and storage constraints.
- mp3: efficient for web
- wav/flac: high fidelity
- opus: streaming friendly
- pcm: uncompressed raw
Character Length Limits
Each API request accepts up to 4096 characters, including spaces and punctuation. For longer training scripts, split the text into segments and join the resulting audio files. This ensures reliable synthesis without timeouts.
Short phrases may take a few seconds to generate, while longer texts process at roughly real-time speed. Plan your batching strategy to maintain smooth production workflows for large course libraries.
- Max 4096 characters per request
- Split longer texts manually
- Join audio files post-generation
Testing and Validation
Testing ensures your Arabic training modules sound natural and professional. Generate samples using the API to evaluate voice clarity before full production.
Use structured feedback to validate audio quality. Adjust parameters like speed and format to match your specific e-learning requirements.
Sample Generation
Send text samples via the API to generate audio files. Use different voices like Sage or Atlas to compare clarity in Arabic script. Limit inputs to 4096 characters per request to ensure consistent processing.
Generate multiple variations to assess how each voice handles instructional content. Select the format that best suits your learning platform.
- Use Sage for calm instruction
- Test Atlas for neutral delivery
- Limit text to 4096 characters
User Feedback Loops
Share generated audio with a small group of learners or instructors. Collect specific feedback on pronunciation and pacing. Use this data to refine voice selection.
Iterate quickly by adjusting the speed parameter between 0.25 and 4.0. Ensure the tone matches the corporate training context.
- Gather learner feedback
- Adjust speed settings
- Refine voice choice
Consistency Checks
Verify that audio quality remains stable across different topics. Check for consistent pronunciation of technical terms in Arabic. Ensure no artifacts appear in the output files.
Validate that the watermark is present and the audio format meets your standards. Confirm that the voice remains neutral and professional throughout the module.
- Check pronunciation consistency
- Verify audio format
- Ensure professional tone
Cost Implications
Pricing is straightforward: $0.01 per 1,000 characters, regardless of whether you select the tts-1 or tts-1-hd model. Failed requests incur no charge, ensuring you only pay for successful synthesis. There are no monthly plans, seats, or minimum spend requirements.
Standard vs HD Pricing
Both tts-1 and tts-1-hd use the same engine and price. Choose based on audio quality needs rather than cost differences. Spaces and punctuation count toward the character limit.
- No tiered pricing based on voice or model
- Same rate for standard and high-definition audio
Volume Discounts
There are no volume discounts. The rate remains constant at $0.01 per 1,000 characters regardless of usage volume. This simplifies budget forecasting for HR agencies producing large-scale training modules.
Trial Usage
Google sign-ups receive $0.50 in trial credit, covering approximately 50,000 characters for seven days. Email-based accounts start with zero balance. You can top up with USDT (TRC-20) with a $5 minimum.
- Trial credit valid for 7 days
- Email accounts do not receive trial credit
Implementation Steps
Begin by generating your API key via the /get-api-key/ page. Use this key in the Authorization header for all requests. Remember that issuing a new key immediately invalidates the previous one.
API Key Setup
Generate your key at /get-api-key/. Include it as Authorization: Bearer tts-... in every request. Note that a new key instantly deactivates the old one.
Voice Parameter Selection
For Arabic training, select voices like Sage for instructional content or Atlas for neutral delivery. Set the language parameter explicitly to avoid incorrect English defaults. Adjust speed between 0.25 and 4.0 to match training pacing.
- Use
tts-1ortts-1-hdmodels. - Set
language: 'ar'for Arabic. - Choose Sage for instructional, Atlas for news-style.
Integration Testing
Send test payloads to POST /v1/audio/speech with short texts. Verify audio formats like MP3 or WAV are returned correctly. Ensure your system handles the 60 requests-per-minute rate limit without queuing delays.
- Test with texts under 4096 characters.
- Verify 200 OK responses for valid inputs.
- Check audio file integrity upon download.

Common Mistakes to Avoid
Selecting the right voice is only half the battle; correct API configuration ensures the output is usable. Misconfigurations lead to wasted credits on unintelligible audio or unexpected billing charges.
This section highlights three frequent errors that corporate training teams make when integrating Arabic TTS into their e-learning workflows.
Ignoring Language Codes
Arabic text in Latin script (transliteration) defaults to English voices, causing unnatural pronunciation. Always set the language parameter to ar for Arabic content, even if the input text uses Latin characters.
Without explicit language codes, the system guesses the language, often resulting in English voices attempting to read Arabic phonemes, which sounds robotic and unprofessional.
- Set
language: 'ar'explicitly. - Use Latin script input if needed, but keep the language tag.
- Verify pronunciation with short samples first.
Overloading Characters
Each API request supports a maximum of 4096 characters. Exceeding this limit causes request failures or truncation, breaking the audio continuity in your training modules.
Split longer scripts into segments of 4096 characters or less before sending them to the API. Join the resulting audio files programmatically to maintain a seamless listening experience.
- Max 4096 characters per request.
- Split long scripts into chunks.
- Concatenate audio files after generation.
Wrong Speed Settings
The speed parameter ranges from 0.25 to 4.0. Setting it too high (e.g., above 2.0) makes Arabic speech difficult to follow, especially for non-native speakers.
For corporate training, a speed between 1.0 and 1.5 is usually optimal. Test different speeds with sample audio to ensure clarity and comprehension for your target audience.
- Range: 0.25–4.0.
- Optimal range: 1.0–1.5 for training.
- Test with native speakers.
Common Mistakes to Avoid
Avoid these pitfalls to ensure high-quality Arabic audio for your corporate training modules.
Proper configuration saves time and credits while maintaining professional standards.
Support and Resources
Access the full API reference at /docs/ to understand request structures, error codes, and integration patterns for HR training pipelines.
For specific questions, contact us via email. We do not offer phone support or live chat, so email is the only channel for assistance.
Documentation Access
The /docs/ page details the POST /v1/audio/speech endpoint, including required fields like model, voice, and input. It explains how to handle rate limits (60 requests per minute) and format options like mp3 or wav.
Review the authentication section to ensure your Authorization: Bearer header is correctly formatted. The documentation also covers the 4096-character limit per request and how to split longer training modules.
Email Support
Send any technical or billing inquiries to our support email. Since we do not provide phone support, email is the primary way to resolve integration issues or clarify voice selection for Arabic content.
Expect a response within one business day. Include your API key (masked) and specific details about your request, such as the voice preset and language code used.
Voice List Endpoint
Use GET /v1/voices to retrieve the complete list of available voices and their properties. This endpoint confirms which voices support Arabic and their associated formats.
The response includes details for Sage, Atlas, and Orion, helping you select the right tone for your e-learning modules without guessing.
- Check voice availability before integrating
- Verify format support per voice
- Confirm Arabic language compatibility
Comparison with Other Solutions
VoxTrain offers eight distinct voice presets, including Sage and Atlas, which are optimized for clear, professional corporate training. Unlike providers that require complex voice cloning or emotion tagging, we provide consistent, standardized voices that ensure uniform audio quality across your entire e-learning catalog.
Voice Variety
Our catalog includes eight presets like Luna and Felix, supporting 23 languages including Arabic. This allows you to maintain brand consistency without the variability of custom voice designs or celebrity imitations.
Pricing Model
We charge $0.01 per 1,000 characters with no monthly plans or seats. Failed requests are free, and you can top up with USTRC-20. This transparent model avoids hidden fees common in subscription-based competitors.
Technical Simplicity
Integration requires only a base URL change in the OpenAI SDK. We support standard formats like MP3 and WAV, with no need for complex SSML parsing or custom lexicons. This reduces development time for HR tech teams.
Final Recommendations
Choosing the right voice depends on your specific training goals. For clarity, Atlas provides a neutral, news-like tone ideal for compliance and technical modules where precision matters. For engagement, Nova offers a bright, upbeat presence that keeps learners attentive during soft skills content. Sage delivers a calm, instructional tone suitable for step-by-step guidance.
For cost efficiency, the standard tts-1 model balances quality and price effectively. Both Atlas and Sage work well with the standard model, ensuring clear audio without the higher cost of HD. Use the free trial to test these voices with your actual Arabic content before scaling production.
Best for Clarity
Atlas is a strong choice for technical documentation and compliance modules where every word must be understood. Its neutral, news-like delivery ensures consistent pronunciation across Arabic dialects, reducing cognitive load for learners processing complex information.
- Use tts-1 model for standard clarity
- Ideal for compliance and technical docs
- Neutral tone reduces listener fatigue
Best for Engagement
Nova’s bright, upbeat personality works best for soft skills training or onboarding videos. The energetic delivery helps maintain learner interest during longer modules, while still maintaining professional clarity for Arabic text.
- Use for onboarding and soft skills
- Bright tone increases attention
- Works well with tts-1
Best for Cost Efficiency
The standard tts-1 model offers the best balance of quality and price at $0.01 per 1,000 characters. Atlas and Sage pair well with this model, providing clear audio without the higher cost of HD. Use the $0.50 trial to validate voice fit before committing to volume.
- Standard tts-1 model is cost-effective
- $0.01 per 1,000 characters
- No hidden fees or subscriptions
Next Steps
Selecting the right voice is only the first step. The next phase involves securing your access credentials and validating the audio output before committing to full-scale production.
With a clear strategy for voice personality and language support, you can move from testing to deployment with confidence.
Get API Key
Navigate to /get-api-key/ to create your account via Google or email. This page displays your unique Bearer token, current balance, and a ready-to-use cURL example for immediate integration.
Ensure you configure your SDK’s base_url to https://api.1930cup.info/v1 to align with the OpenAI-compatible endpoint structure.
- Sign up with Google or email + password
- Retrieve your Bearer token from the dashboard
- Verify your balance and available trial credit
Generate Samples
Use the /v1/audio/speech endpoint with your chosen voice (Sage, Atlas, or Orion) and the 'ar' language code. Send short text segments to verify pronunciation clarity and tone suitability for your specific training modules.
Review the generated audio files to ensure they meet your quality standards before scaling up.
- Test Sage for instructional content
- Test Atlas for neutral narration
- Test Orion for documentary-style modules
Start Production
Once validated, integrate the API into your LMS or content pipeline. Remember to split texts over 4096 characters and join the resulting audio files.
Monitor your usage via the dashboard to manage your $0.01 per 1,000 characters cost efficiently.
- Split long texts into 4096-character chunks
- Join audio files for final output
- Track usage and balance in real-time
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.
Sources and further reading
- Wikipedia
- MDN: Web Speech API
- TRON developer hub: TRC-20
- FFmpeg documentation
- OpenAI API reference: create speech
- IETF RFC 7845: Ogg Opus
- Creative Commons Attribution 4.0
- W3C WAI: making audio and video accessible
- LibriTTS-R paper (arXiv 2305.18802)
- ISO 639 language codes
- LibriTTS-R corpus (OpenSLR 141)
- LibriVox public-domain audiobooks
