Frequently asked questions
What are the best TTS voices for education?
Q: Which TTS voices are suitable for education? A: Sage offers a calm, instructional tone ideal for lectures, while Aria provides a warm, clear narrator style suitable for general course content. Both are effective for delivering educational material clearly without distraction, supporting diverse learning styles through distinct vocal characteristics.
Which voice is best for reading novels?
Luna is designed for soft, storytelling delivery, making it well-suited for narrative content. Orion provides a deep, documentary style that adds gravity to longer texts. These presets help maintain listener engagement during extended reading sessions in online courses. Note that any voice can speak any of the 23 supported languages, so selection depends on your specific tonal preference rather than a single "best" option.
Are male voices better than female voices?
Neither is objectively better; they serve different tonal needs. Female voices include Aria, Nova, Luna, and Sage. Male voices include Felix, Milo, Atlas, and Orion. Selection depends on whether you prefer a warm narrator, upbeat energy, or calm instruction for your specific course material.
Can voices speak multiple languages?
Yes. Any of the eight voices can speak any of the 23 supported languages, including English, Spanish, French, German, Japanese, and Arabic. You can specify the language via an ISO code or let the system guess it from the script.
Is there a cost difference between voices?
No. All eight voices use the same engine and pricing structure. Whether you select Aria, Felix, or any other preset, the cost remains $0.01 per 1,000 characters, regardless of the voice chosen or the language spoken.
How do I choose a voice for my course?
Match the voice to your content type. Use Sage for instructional modules, Luna for storytelling, or Atlas for neutral news-style updates. Test samples using the API to ensure the tone aligns with your course’s educational goals and audience expectations.
Can I adjust voice speed?
Yes. You can set the speed parameter between 0.25 and 4.0. This allows you to slow down complex explanations for clarity or speed up repetitive content, ensuring the audio pacing fits your learners’ preferences and course structure effectively.
What is the difference between tts-1 and tts-1-hd?
Both models use the same engine and price. The distinction lies in the model identifier; both use the same engine and price. However, since pricing is identical, you can choose either model based on your specific audio quality preferences without extra cost.
Are the voices synthetic or human?
The voices are synthetic, generated from licensed recordings of human actors. Each audio file contains an inaudible watermark marking it as synthetic speech. They are not direct recordings but AI-generated interpretations based on the original licensed voice performances.
Do voices have accents?
The voices are designed to speak clearly in their supported languages without strong regional accents. If you specify a language code, the voice adapts to that language’s standard pronunciation. Default guessing assumes English for Latin script text if no language is set.
Can I use voices in commercial courses?
Yes. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. You can generate audio for commercial online courses and LMS platforms without additional licensing fees beyond the standard per-character API usage cost.
How do I test voice quality?
Send a short text sample via the POST /v1/audio/speech endpoint. Use the free trial credit of $0.50 to generate audio files in formats like mp3 or wav. Review the output to ensure the tone and clarity meet your course standards.
Is there a way to preview voices?
You can generate short audio samples by sending requests to the API. Use the /v1/voices endpoint to list available voices and their attributes. Generate small text snippets to hear each voice’s tone, speed, and clarity before committing to full course generation.
What is the character limit per request?
Each API request accepts up to 4096 characters in the input field. For longer texts, split the content into chunks of this size, generate audio for each part, and join the resulting files to create a complete course module audio track.
How do I handle voice changes in API updates?
The API uses fixed model identifiers like tts-1 and tts-1-hd. Voice presets are fixed. You can monitor the /v1/voices endpoint for any changes to available voices or their attributes, ensuring your integration remains compatible with current offerings.
Are there any voice limitations?
Voices do not support SSML, emotion tags, or custom pronunciation lexicons. They cannot perform voice cloning or mimic specific celebrities. Each voice has a fixed tone profile, so you cannot adjust pitch or intonation beyond the standard speed parameter range.
Can I mix voices in one audio file?
No. Each API request generates audio using a single voice preset. To mix voices, you must generate separate audio files for each voice segment and join them using external audio editing tools or your LMS’s media processing capabilities.
Do voices support pauses or emphasis?
The API does not support SSML tags for explicit pauses or emphasis. Punctuation in the input text influences natural phrasing, but you cannot programmatically insert specific pause durations or stress words. Control pacing primarily through text structure and speed adjustments.
Is there a limit to the number of voices?
There are exactly eight voice presets available: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion. You can use any of these voices in your requests without limit, but new voices are not added.
How do I get started with voice selection?
Sign up for an account to receive your API key. Use the /v1/voices endpoint to view available options. Generate test audio using the free trial credit to evaluate each voice’s tone, clarity, and suitability for your specific educational content types.
Overview of Voice Options
Choosing the right voice is critical for maintaining student engagement and ensuring clarity in digital learning environments. VoiceLoom provides eight distinct presets, each optimized for different educational contexts, from immersive storytelling to structured instructional delivery. The API supports 23 languages, allowing educators to reach diverse global audiences without compromising audio quality.
The selection process balances auditory comfort with content retention. Whether narrating a dense lecture or guiding a interactive module, the available voices offer distinct tonal qualities that align with specific pedagogical goals. Understanding these nuances helps product teams integrate audio that enhances, rather than distracts from, the learning material.
Female voices overview
The female voice options cater to varied instructional styles. Aria provides a warm, clear narration ideal for standard course content. Nova offers a bright, upbeat tone suitable for engaging introductions. Luna delivers a soft, storytelling quality perfect for narrative-driven modules.
These voices are designed for clarity and approachability. They support the 23 available languages and can be adjusted via speed settings to match listener preferences. Each voice carries an inaudible watermark, ensuring synthetic origin transparency without affecting the listening experience.
- Aria: Warm, clear narrator for standard content.
- Nova: Bright, upbeat tone for engaging intros.
- Luna: Soft, storytelling quality for narratives.
Male voices overview
Male voices in the API cover a spectrum from energetic to neutral. Felix is energetic and suited for promotional or high-energy segments. Milo provides a friendly, conversational tone, making complex topics feel approachable. Atlas remains neutral, ideal for news-style updates or factual delivery.
These voices ensure consistent delivery across different content types. Like the female options, they support multiple languages and speed adjustments. The consistent pricing model means switching between these presets does not incur additional costs, allowing flexible audio strategy.
- Felix: Energetic for promo or high-energy segments.
- Milo: Friendly, conversational for approachable explanations.
- Atlas: Neutral for news or factual updates.
Voice characteristics
All voices are based on licensed recordings from LibriTTS-R and LibriVox public-domain audiobooks, ensuring a natural, human-like quality without the uncanny valley effect. The API processes requests sequentially, so longer texts may queue, but typically generate at roughly real-time speed.
Key characteristics include support for 23 languages, speed adjustment between 0.25x and 4.0x, and multiple output formats like MP3 and WAV. Each generated file is marked with an inaudible watermark, identifying it as synthetic speech for transparency and compliance purposes.
- Licensed recordings from LibriTTS-R and LibriVox.
- Support for 23 languages with auto-detection.
- Speed adjustment from 0.25x to 4.0x.
- Inaudible watermark for synthetic identification.
image]:
Female Voices for Education
Choosing the right female voice for LMS content directly influences learner retention and comprehension. VoiceLoom offers three distinct presets designed for educational clarity and engagement, each derived from licensed recordings to ensure high audio fidelity without licensing overhead.
These voices support 23 languages and adapt seamlessly to course scripts, from concise quizzes to extended lectures. The consistent $0.01 per 1,000 characters pricing applies equally across all female presets, making budgeting predictable for product teams scaling course libraries.
Aria: Warm and clear
Aria serves as an ideal narrator for general educational content. Her warm, clear tone reduces cognitive load, making her suitable for introductory modules or complex explanations where listener fatigue is a concern. She provides a neutral yet engaging backdrop for text-heavy courses.
Use Aria for standard reading passages, vocabulary definitions, or general knowledge summaries. Her clarity ensures that non-native speakers can follow along without straining to distinguish phonemes.
- Ideal for: General lectures, definitions, and standard reading passages
- Best for: Reducing listener fatigue in long-form content
- Language support: Works across all 23 supported languages
Nova: Bright and upbeat
Nova brings energy to instructional material, particularly effective for short, engaging modules or interactive exercises. Her bright, upbeat tone keeps learners alert and can increase completion rates for quick quizzes or promotional course updates.
Deploy Nova for micro-learning snippets, announcements, or dynamic content that requires a lively delivery. Her energetic presence helps maintain attention during brief, high-density information bursts.
- Ideal for: Micro-learning, quizzes, and course announcements
- Best for: Maintaining learner alertness during short sessions
- Language support: Effective in English and other Latin-script languages
Luna: Soft and storytelling
Luna excels in narrative-driven courses, such as history, literature, or case studies. Her soft, storytelling voice creates an immersive atmosphere, drawing learners into the content without overpowering the message. This makes her perfect for audiobook-style modules.
Use Luna for historical narratives, biographical case studies, or any content that benefits from a gentle, engaging tone. Her style supports deep focus by minimizing auditory distraction while maintaining emotional connection to the material.
- Ideal for: Case studies, historical narratives, and literature
- Best for: Immersive, narrative-driven learning experiences
- Language support: Adapts well to emotional nuance in supported languages
Male Voices for Education
Male voices offer distinct tonal qualities for educational content, ranging from high-energy promo styles to neutral news delivery. Each preset serves a specific pedagogical purpose, allowing product teams to match audio output with course material density and target audience expectations.
Felix: Energetic and promo
Felix delivers an energetic tone ideal for promotional course trailers or introductory modules. His clear, upbeat delivery keeps learners engaged during short, high-impact segments where attention retention is critical.
- Best for video intros and course teasers
- High energy maintains viewer interest
- Clear articulation for marketing-style content
Milo: Friendly and conversational
Milo provides a friendly, conversational tone suitable for tutorial walkthroughs and explanatory content. His natural pacing helps simplify complex concepts, making him ideal for step-by-step instructional videos.
- Ideal for tutorial and how-to guides
- Conversational tone reduces cognitive load
- Supports clear concept explanation
Atlas: Neutral and news
Atlas offers a neutral, news-style delivery for factual content or data-heavy modules. His consistent tone ensures unbiased presentation of information, making it suitable for academic or professional training materials.
- Neutral tone for unbiased content
- Suitable for news-style updates
- Consistent delivery for data modules
Specialized Voices
Beyond standard narration, specific voices optimize distinct educational contexts. Sage delivers calm, instructional clarity ideal for complex tutorials, while Orion provides a deep, documentary tone suited for immersive historical or scientific content.
These specialized options allow course creators to match audio personality directly to subject matter, enhancing retention without requiring voice actor hires.
Sage: Calm and instructional
Sage is designed for instructional clarity. Its calm tone reduces cognitive load during dense explanations, making it suitable for step-by-step tutorials or technical documentation.
Use Sage for module introductions where clarity matters more than energy. The voice remains neutral enough to avoid distracting from the core material.
- Best for: Technical how-tos, software tutorials, procedural guides
- Tone: Calm, steady, instructional
Orion: Deep and documentary
Orion offers a deep, authoritative voice ideal for narrative-driven content. Its documentary style engages learners in historical, scientific, or biographical courses.
Use Orion for case studies or long-form reading content where a consistent, immersive tone sustains attention over longer durations.
- Best for: History lessons, science documentaries, case studies
- Tone: Deep, authoritative, immersive
Use cases for each
Pair Sage with high-density information modules and Orion with narrative-heavy sections. This contrast can help structure course pacing, signaling shifts between learning modes.
Both voices support all 23 languages, allowing consistent branding across international courses without re-recording content.
Voice Selection Criteria
Selecting the right voice for educational content requires balancing auditory clarity with emotional resonance. Product teams must evaluate how well a voice articulates complex course material without causing listener fatigue during extended study sessions.
The API’s eight presets offer distinct characteristics, from warm narrators to neutral news readers. Choosing wisely ensures the audio complements the text rather than distracting from it.
Clarity and articulation
Clear articulation is critical for comprehension, especially in language learning or technical courses. Voices like Aria and Sage provide crisp enunciation, ensuring that every word is distinct. This reduces the cognitive load on learners, allowing them to focus on the content rather than deciphering the speech.
For dense material, a neutral tone like Atlas helps maintain consistency. The API’s standard model delivers reliable pronunciation across 23 languages, supporting diverse student bodies without requiring multiple voice profiles.
- Prioritize voices with consistent enunciation for technical terms.
- Test audio at varying speeds (0.25–4.0) to ensure clarity remains intact.
- Use Sage for instructional guides to maintain a calm, clear delivery.
Tone and mood
The emotional tone of the voice should align with the subject matter. A warm voice like Aria suits humanities, while the energetic Felix works well for promotional course trailers. Matching tone to content enhances retention and keeps learners engaged.
Avoid mismatched tones, such as using a deep documentary voice (Orion) for a light introductory module. Consistency in mood helps establish a predictable learning environment, which is crucial for user experience in LMS platforms.
- Match voice personality to course topic (e.g., Nova for upbeat modules).
- Maintain consistent tone across a course to avoid jarring transitions.
- Use Luna for storytelling modules to enhance narrative flow.
Audience engagement
Engagement is driven by how relatable the voice feels. Conversational tones like Milo foster a sense of direct instruction, making learners feel guided rather than lectured. This approach is particularly effective for self-paced online courses.
A/B testing different voices with sample groups can reveal preferences. Since the API returns audio files quickly, teams can iterate on voice selection based on actual listener feedback, optimizing for completion rates.
- Use Milo for conversational, student-facing content.
- Conduct A/B tests with different voices to measure engagement.
- Consider audience demographics when selecting between warm or neutral tones.
Technical Specifications
The API uses standard OpenAI-compatible request shapes, allowing product teams to integrate VoiceLoom into existing LMS workflows with minimal code changes. By pointing the SDK’s base_url to https://api.nbcasino.info/v1, teams can utilize familiar libraries to send text and receive audio files directly within their course modules.
Voice ID usage
Each voice is referenced by a simple string ID (e.g., "aria", "nova") in the voice field of the JSON body. This consistent naming convention ensures that audio assets remain linked to specific narrative styles across different course modules, allowing for reliable audio regeneration if assets are cleared.
- Use preset IDs like "aria" or "milo" directly in the request body.
- IDs are case-sensitive and match the voice list exactly.
- No complex JSON structures required for voice selection.
Language support
The API supports 23 languages, including English, Spanish, German, and Japanese. If the language parameter is omitted, the system guesses the language from the script, defaulting to English for Latin text. For non-English Latin-script content, explicitly setting the ISO code ensures accurate pronunciation and avoids misinterpretation.
- Supports 23 languages including Arabic, Chinese, and Russian.
- Explicit
languagecode recommended for non-English Latin text. - Automatic guessing defaults to English if unspecified.
Speed adjustment
The speed parameter accepts values between 0.25 and 4.0, allowing instructors to adjust playback speed for clarity or emphasis. Slower speeds aid comprehension for complex academic material, while faster speeds suit review sessions. The API processes requests sequentially, so speed adjustments do not affect queue latency under normal load.
- Range: 0.25 to 4.0.
- Adjustable per request for different content types.
- Does not incur additional cost beyond character count.
Pricing Impact
The pricing model is straightforward: $10 per million characters, regardless of the voice or quality tier selected. This uniformity means you can experiment with different voices for various course modules without worrying about variable costs, making budget forecasting for LMS audio production simple and predictable.
Uniform pricing across voices
Every voice, from Aria to Orion, costs the same per character. This consistency allows product teams to allocate budget based on content needs rather than voice prestige, ensuring that high-quality audio remains accessible for all course segments.
- Same rate for all 8 voices
- Predictable cost per character
No extra cost for HD
The tts-1-hd model uses the same pricing as the standard tts-1 engine. You can upgrade to higher fidelity for critical lectures or language learning modules without incurring additional fees, allowing you to optimize audio quality based on content importance alone.
- tts-1 and tts-1-hd share the same rate
- No premium for high definition
Cost-effective options
With a free trial of $0.50 and a low minimum top-up of $5 via USDT on the TRON network (TRC-20), testing is low-risk. The per-character billing ensures you only pay for what you use, making it ideal for small course batches or pilot programs without monthly commitments.
- $0.50 free trial
- Minimum $5 top-up
- No monthly subscriptions

User Experience
User experience hinges on how well the synthetic voice supports learning objectives. With eight distinct presets, product teams can match audio tone to content type, ensuring that narration feels natural rather than robotic.
Listener preference
Learners respond differently to voice characteristics. Aria offers a warm, clear narrator style ideal for complex explanations, while Milo provides a friendly, conversational tone better suited for casual tutorials.
- Match voice personality to course tone
- Avoid monotony by alternating voices
Accessibility compliance
Clear, consistent audio helps meet accessibility standards for visually impaired users. The API supports multiple languages, allowing localized audio for diverse student bases without requiring multiple voice actors.
- Supports 23 languages
- Consistent pronunciation across texts
Engagement metrics
Adjustable speed (0.25–4.0) lets users control playback to fit their pace. Faster speeds help review known material; slower speeds aid comprehension of difficult concepts.
- Customizable playback speed
- No extra cost for HD quality
Implementation Tips
Selecting the right voice is critical for educational content, but technical implementation requires precise configuration. The API accepts a speed parameter between 0.25 and 4.0, allowing you to adjust narration pace for complex material or quick reviews. Since requests are processed sequentially, long texts must be split into chunks of up to 4096 characters to ensure consistent generation without queue delays exceeding reasonable limits.
Audio files are returned directly in the response, simplifying storage workflows. You can generate audio on demand or pre-render lessons during content creation. Using the OpenAI SDK with a adjusted base_url ensures compatibility with existing LMS integrations, reducing development overhead for engineering teams.
Testing voices
Before full deployment, test each voice preset with your specific course material. Voices like Aria and Luna perform differently with narrative text compared to instructional content from Sage or Atlas. Verify that the language detection works correctly for non-English Latin scripts by explicitly setting the language field.
Check audio quality across different formats. While MP3 is standard, WAV or FLAC may be preferred for archival purposes. Ensure the 4096-character limit does not truncate meaningful sentences, as this affects the listener's experience.
- Test with varied text lengths to observe splitting behavior
- Verify language detection accuracy for mixed-script content
- Compare audio clarity across MP3, WAV, and FLAC
User feedback
Collect direct feedback from learners regarding voice preference and comprehension. Since the API does not offer custom voice cloning or emotion tags, rely on the eight preset voices to match your brand tone. Monitor if listeners fatigue from repetitive voices during long sessions.
Use the free trial credit to conduct small-scale tests with actual users. This helps identify which voices resonate best before committing to larger volume usage. Track characters consumed to estimate budget impact accurately.
- Survey learners on voice preference and clarity
- Monitor fatigue levels during extended listening sessions
- Use trial credit for initial user validation
A/B testing
Run A/B tests comparing different voices for the same lesson. For example, compare Nova’s upbeat tone against Milo’s conversational style for introductory modules. Measure engagement metrics such as completion rates and time spent on each audio version.
Use these insights to standardize voice choices for specific content types. News-style content may suit Atlas, while storytelling benefits from Luna. Document these preferences to maintain consistency across your course catalog.
Ensure consistent character counting for accurate cost comparison between test variants. This data helps refine your production workflow and optimize spending on high-performing audio assets.
- Compare voice performance across different lesson types
- Track completion rates and engagement metrics
- Standardize voice choices based on data insights
Comparative Analysis
Comparing voice options helps product teams align audio tone with specific course modules. While all voices share the same engine and pricing, their distinct characteristics influence learner retention and comprehension differently.
Voice comparison chart
The table below contrasts the eight available presets by tone and ideal use case, ensuring you match the voice to the content type.
- Aria: Warm narrator for general lessons.
- Nova: Upbeat for interactive quizzes.
- Luna: Soft storytelling for narratives.
- Felix: Energetic for promotional content.
- Milo: Conversational for tutorials.
- Atlas: Neutral for news updates.
- Sage: Calm for instructional guides.
- Orion: Deep for documentary styles.
Pros and cons
Using distinct voices prevents listener fatigue during long courses, but managing multiple voice IDs adds slight complexity to your API integration. All voices support all 23 languages, ensuring consistency across global learners without extra cost.
- Pros: Enhanced engagement through variety; consistent pricing across all models.
Recommendations
For accessibility-heavy modules, use Aria or Sage for clarity. For immersive storytelling, Luna or Orion work best. Always test with sample text to verify articulation in your target language before full rollout.
Future-Proofing
As your course library grows, the API handles volume efficiently. Requests are processed sequentially, meaning longer texts may queue under heavy load, but the system scales to meet demand without complex infrastructure changes.
Pricing remains uniform regardless of voice choice, ensuring predictable costs as you expand your content library without hidden fees for higher quality models.
Scalability
The API processes requests sequentially, so under heavy load, a request may wait in a queue. This ensures stable performance but requires planning for peak enrollment periods.
- Sequential processing ensures stability
- No server management required
New voice additions
Once available, they appear in the /v1/voices endpoint, allowing you to offer fresh options without code changes.
API updates
The endpoint structure remains stable, compatible with standard OpenAI SDKs. Updates are backward-compatible, minimizing disruption to your existing LMS integrations.
Conclusion
VoiceLoom provides a straightforward, cost-effective solution for generating course audio at scale. With uniform pricing regardless of voice choice and support for 23 languages, it simplifies the technical side of accessibility and immersion without hidden fees or complex infrastructure.
Product teams can integrate the API directly into their LMS workflows, leveraging existing OpenAI SDK compatibility for rapid deployment. The clear distinction between voice presets allows for consistent branding across diverse educational content types.
Final recommendations
Prioritize Aria or Luna for narrative-heavy modules and Atlas for factual content. Test speed adjustments to match listener comprehension levels, keeping generation times reasonable for longer texts.
- Use Aria for general narration
- Select Atlas for news-style updates
- Adjust speed between 1.0–2.0 for clarity
Next steps
Sign up for the $0.50 trial to generate initial audio samples. Evaluate character limits and splitting strategies for your longest course modules before full integration.
- Claim trial credit
- Test 4096-character limits
- Integrate via API base
Resources
Refer to the documentation for endpoint details and the pricing page for current rates. Support is available via email for any integration questions.
- /docs/
- /pricing/
- /get-api-key/
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Choosing the right female voice for LMS content directly influences learner retention and comprehension.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- Creative Commons Attribution 4.0
- W3C WAI: making audio and video accessible
- Tether (USDT)
- LibriTTS-R paper (arXiv 2305.18802)
- LibriTTS-R corpus (OpenSLR 141)
- openai-python on GitHub
- ISO 639 language codes
- FFmpeg documentation
- Python documentation
- OpenAI API reference: create speech
- TRON developer hub: TRC-20
