Frequently asked questions
Which voice is best for French?
All eight voices support French. Choose Aria for a warm, clear narrator style or Luna for soft storytelling. Set the language parameter to 'fr' to ensure correct pronunciation, as the system guesses the language from the script by default. This ensures accurate delivery for your educational materials without needing to switch voices.
Are the voices human?
The voices are based on licensed recordings from LibriTTS-R, which uses CC BY 4.0 data from LibriVox public-domain audiobook readings. While they originate from human speakers, the generated audio files carry an inaudible watermark that marks them as synthetic speech, distinguishing them from direct human recordings.
Can I mix voices?
The API generates one audio file per request using a single voice preset. To mix voices, you must send multiple requests with different voice parameters and then join the resulting audio files yourself. There is no built-in feature to blend voices within a single API call.
Do voices sound robotic?
The voices are derived from real audiobook readings, offering natural intonation. However, every generated file carries an inaudible watermark marking it as synthetic speech. The tone is clear and consistent, suitable for educational materials, but listeners may detect the synthetic nature compared to raw human recordings.
Which voice is calmest?
Sage is described as calm and instructional, making it ideal for steady reading or lessons. Luna is also soft and suited for storytelling. For French language learning, Sage provides a neutral, relaxed pace that helps students focus on pronunciation without energetic distractions.
Which voice is most energetic?
Felix is described as energetic and promo-oriented, while Nova is bright and upbeat. If you need lively audio for engaging students or dynamic language exercises, Felix or Nova are the best choices among the eight presets to convey high energy and enthusiasm.
Can I use voices for kids?
All eight voices support French and can be used for children's content. You can adjust the speed between 0.25 and 4.0 to make the speech clearer or slower for young learners. The voices are clear and distinct, suitable for educational reading materials and language tools.
Do voices support emojis?
The API accepts text input up to 4096 characters. Emojis are treated as characters within that limit. However, the system does not support SSML or custom pronunciation lexicons, so emojis may not be spoken aloud or interpreted as specific sounds.
Is there a female/male split?
Yes, there are four female voices (Aria, Nova, Luna, Sage) and four male voices (Felix, Milo, Atlas, Orion). You can choose any voice regardless of the text content. All voices can speak any of the 23 supported languages, including French, with no gender restriction on language.
Can I change speed per voice?
You can adjust the speed parameter between 0.25 and 4.0 for any voice. This allows you to slow down speech for better comprehension or speed it up for quick reviews. The speed setting applies to the specific request and does not permanently change the voice preset.
Are voices consistent?
The voices are based on licensed recordings from LibriVox, ensuring a consistent synthetic style. Each request generates audio using the same engine parameters. While the tone is consistent, the exact output may vary slightly due to queue processing, but the voice character remains stable across requests.
Do voices have accents?
The voices are based on recordings from LibriVox, which features native speakers. When you set the language to 'fr', the voices will speak with appropriate French pronunciation. If the language is not set, the system guesses it, so setting it explicitly ensures the correct accent for French.
Which voice is fastest?
Q: Which voice is fastest? A: All eight voices generate audio at the same speed, roughly real-time (25 characters per second). The speed parameter (0.25–4.0) adjusts playback speed, not generation time. Processing happens sequentially, so generation time depends on queue load, not the specific voice you choose.
Can I clone a voice?
No, custom voice cloning is not offered. You cannot upload your own voice or design a new voice. You must choose from the eight available presets: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, or Orion. These are the only voices available for your API requests.
Criteria for Choosing Voices
Selecting the right voice for French language learning requires balancing linguistic accuracy with student engagement. A clear, precise pronunciation helps learners distinguish subtle phonetic nuances, while an engaging tone maintains attention during repetitive drills or reading exercises. The voice must sound natural yet distinct enough to aid in auditory discrimination tasks.
Clarity and Pronunciation
Educators need voices that articulate French phonemes accurately without excessive accent drift. The API handles French natively, ensuring consistent stress patterns and vowel quality. This reliability allows methodologists to trust the audio as a standard reference for students practicing listening comprehension.
- Consistent French phoneme articulation
- Clear enunciation for difficult sounds
- No artificial accent drift
Tone and Engagement
Monotone audio can disengage students quickly. Voices like Nova or Aria offer varied intonation that mimics natural speech rhythms, keeping learners interested. A warmer tone encourages repeated listening, which is crucial for mastering new vocabulary and sentence structures.
- Natural intonation patterns
- Varied emotional warmth
- Encourages repeated listening
Suitability for Age Group
Younger students respond better to brighter, faster-paced voices like Nova, while older learners may prefer the calm, instructional tone of Sage. Matching voice speed (0.25–4.0) to the student’s level ensures the audio is neither too rushed nor too slow for effective processing.
- Adjustable speed (0.25–4.0)
- Age-appropriate voice selection
- Customizable for learning levels
Top Female Voices
Selecting the right voice is essential for maintaining student engagement during language learning. Our female voices offer distinct characteristics tailored to different educational contexts, from clear pronunciation drills to immersive storytelling sessions.
Aria: Warm and Clear
Aria delivers a warm, clear narrator tone that excels in vocabulary exercises. Her consistent pacing helps students distinguish individual phonemes, making her ideal for beginners focusing on accurate pronunciation and listening comprehension.
- Ideal for vocabulary drills
- Clear enunciation for beginners
- Warm, non-distracting tone
Nova: Bright and Upbeat
Nova brings a bright, upbeat energy perfect for keeping younger learners engaged. Her lively delivery makes repetitive practice feel less monotonous, encouraging students to stay attentive during grammar lessons or interactive dialogues.
- High energy for engagement
- Great for interactive drills
- Keeps attention focused
Luna: Soft Storytelling
Luna offers a soft, storytelling voice suited for reading passages and narrative texts. Her gentle tone creates a calm learning environment, reducing anxiety for students who need a soothing audio experience during extended reading sessions.
- Soft, calming tone
- Perfect for long texts
- Reduces listening fatigue
Top Male Voices
Male voices offer distinct auditory profiles for diverse educational needs, from energetic announcements to neutral news delivery. Each preset serves a specific pedagogical purpose, ensuring audio content matches the lesson's tone without requiring technical expertise from teachers.
Felix: Energetic Promo
Felix brings an energetic, promo-style energy ideal for capturing student attention during transitions or introductions. His clear, upbeat delivery works well for short announcements or motivational segments in language classes.
- Ideal for short attention spans
- Clear calls to action
- High energy, engaging tone
Milo: Friendly Conversational
Milo provides a friendly, conversational tone perfect for dialogues and interactive practice. His natural flow makes audio lessons feel like a personal tutor, encouraging students to engage with spoken French.
- Natural dialogue flow
- Approachable instructional style
- Great for conversational exercises
Atlas: Neutral News
Atlas delivers a neutral, news-style voice suitable for factual content and current events. His consistent, objective tone ensures clarity and reduces distraction, making him ideal for reading comprehension materials.
- Neutral, professional delivery
- Objective tone for facts
- Clear pronunciation focus
Voice for Storytelling
Luna offers a soft, storytelling voice ideal for bedtime stories or immersive reading sessions. Her clear enunciation helps students focus on vocabulary and rhythm without distraction.
Luna for Bedtime Stories
Use Luna to generate calming audio for pre-sleep reading or quiet classroom time. Her gentle pacing supports relaxation while maintaining educational value.
- Best for ages 6–12
- Ideal for quiet hours
- Supports language immersion
Engaging Narration
Luna’s tone keeps learners engaged without overstimulation. She handles narrative arcs smoothly, making complex texts easier to follow for developing readers.
Emotional Tone
While she lacks explicit emotion tags, Luna’s natural inflection conveys warmth and care. This subtle expressiveness helps students connect emotionally with stories.
- No SSML needed
- Natural inflection
- Consistent delivery
Voice for Instruction
educators need audio that guides students through exercises without distraction. Sage delivers a calm, instructional voice perfect for lesson overviews and step-by-step directions. Her clear articulation ensures learners focus on the content rather than the speaker’s style.
Sage for Lessons
Sage’s neutral yet warm tone works well for introducing new vocabulary or grammar rules. She maintains a steady pace, making complex instructions easier to digest for students of varying levels.
Clear Instructions
Use Sage for reading comprehension tasks or listening exercises. Her consistent delivery helps learners identify key information without emotional bias interfering with the educational material.
- Step-by-step directions
- Grammar rule explanations
- Vocabulary definitions
Professional Tone
Unlike conversational voices, Sage avoids excessive energy, keeping the focus on the text. This professionalism suits formal school settings and standardized test preparation materials.
Voice for News and Facts
The API delivers audio files quickly, allowing teachers to generate fresh content for current events lessons without waiting for manual processing. This reliability ensures that classroom materials remain relevant and up-to-date.
Atlas for Current Events
Atlas provides a neutral, news-style voice ideal for reading daily headlines or summarizing recent history. Its clear enunciation helps students focus on content rather than vocal quirks, making it perfect for standardized reading materials.
- Ideal for reading daily headlines
- Clear enunciation for focus
- Standardized reading materials
Neutral Delivery
Unlike conversational voices, Atlas avoids emotional inflection, presenting facts objectively. This neutrality prevents bias in educational content, ensuring that historical or scientific facts are delivered with consistent clarity and professional tone.
Objective Tone
The objective tone supports critical thinking by removing subjective vocal cues. Students can analyze the text’s meaning based on words alone, which is essential for language learning exercises requiring precise comprehension of factual data.
Voice for Conversational Practice
For language educators, conversational practice bridges the gap between reading text and hearing natural speech. Milo provides a friendly, conversational voice that mimics the rhythm of everyday dialogue, making it ideal for learners practicing pronunciation and intonation.
Milo for Dialogues
Milo’s warm tone reduces the stiffness often found in synthetic speech. Teachers can generate short dialogue exchanges or monologues that feel authentic, helping students focus on comprehension rather than adjusting to artificial cadences.
- Ideal for role-play exercises and pair work audio tracks
- Supports clear enunciation for beginner and intermediate learners
Natural Flow
The API processes text efficiently, allowing you to create varied dialogue snippets without significant delay. With speed adjustable between 0.25 and 4.0, educators can slow down Milo’s speech for younger students or speed it up for advanced listening practice.
- Adjustable speed ensures accessibility for different proficiency levels
- Consistent voice quality across multiple dialogue segments
Interactive Style
Use Milo to create audio prompts for interactive lessons. Since the API supports multiple formats like mp3 and wav, these audio files integrate easily into learning management systems or classroom presentations. The lack of SSML keeps the workflow simple for methodologists who prioritize content over technical configuration.
- Seamless integration into standard classroom audio workflows
- Clear, distinct speech patterns aid listener comprehension

Voice for Announcements
Felix brings necessary energy to school announcements, ensuring students notice important updates. His energetic delivery cuts through background noise, making daily schedules, event reminders, and emergency alerts impossible to miss. This voice transforms routine broadcasts into engaging moments that capture attention immediately.
Felix for Energies
Use Felix for high-priority messages like lunch changes or assembly reminders. His upbeat tone keeps the atmosphere positive and lively, even when delivering standard administrative information.
Attention Grabbing
The distinct vocal profile of Felix stands out against other voices. This uniqueness helps students instantly recognize that an announcement requires their full attention, reducing distractions in hallways or classrooms.
Clear Calls to Action
Felix excels at delivering clear instructions. Whether telling students where to line up or reminding them of homework deadlines, his clear enunciation ensures the message is understood correctly the first time, minimizing confusion and repeated questions.
Voice for Documentaries
Orion provides a deep, resonant voice ideal for documentary-style educational content. His serious tone brings weight and authority to historical narratives, making complex topics feel significant and engaging for students.
Orion for Deep Narration
Orion’s distinctively deep register offers a professional documentary feel, perfect for long-form educational materials. The voice remains steady and clear, ensuring that historical facts are delivered with the gravity they deserve without distracting from the content.
Serious Tone
The serious tone helps maintain focus during dense informational segments. Unlike lighter voices, Orion’s delivery emphasizes the importance of the subject matter, keeping learners attentive during detailed explanations of scientific or historical events.
Historical Context
Use Orion when teaching history or science documentaries. His narration style provides a consistent, authoritative backdrop that helps students distinguish between casual conversation and formal educational content, enhancing their understanding of context.
Remember that Orion is one of eight presets available via the API. You can adjust the speed between 0.25 and 4.0 to ensure the narration matches your classroom’s pacing needs.
Technical Considerations
Enhancing audio accessibility requires precise technical configuration. By adjusting the speed parameter between 0.25 and 4.0, educators can tailor speech rate to student comprehension levels, slowing down complex texts or accelerating familiar content for review.
Speed Adjustment
Set speeds lower than 1.0 for beginners or higher for advanced learners. This flexibility ensures the audio matches the cognitive load of the lesson, keeping students engaged without overwhelming them with rapid delivery.
- Range: 0.25 to 4.0
- Adjust per difficulty level
Language Setting
Explicitly define the language using ISO codes to avoid guess errors, especially for non-English Latin scripts. While the API guesses from script, forcing French ensures accurate pronunciation for words that might otherwise default to English.
- Use ISO codes
- Prevents default English errors
Format Selection
Choose MP3 for web players, WAV for high-quality editing, or FLAC for lossless archives. Selecting the right format balances bandwidth efficiency with audio fidelity, ensuring smooth playback on classroom devices and storage systems.
- MP3 for web
- WAV for editing
- FLAC for archives
Testing Your Choices
Begin by generating short samples using your $0.50 free trial credit, which provides roughly 50,000 characters of usage valid for seven days after signing in with Google. This allows you to test voice clarity without financial commitment, ensuring the audio meets your educational standards before purchasing larger blocks of characters.
Use Free Trial
Leverage the trial credit to experiment with different speed settings between 0.25 and 4.0, optimizing the listening pace for various student age groups. Since requests are processed sequentially, shorter test phrases generate quickly, allowing rapid iteration without waiting in long queues.
- Sign in via Google to instantly receive $0.50 credit.
- Test up to 50,000 characters across multiple voices.
- Adjust speed to match student comprehension levels.
Compare Audio Quality
Download samples in mp3, wav, opus, flac, or pcm formats to evaluate fidelity across different devices used in classrooms. Each generated file contains an inaudible watermark marking it as synthetic speech, ensuring transparency while you assess naturalness and clarity for language learning materials.
- Test output formats like wav for high quality or mp3 for easy sharing.
- Verify clarity on standard classroom speakers and student headphones.
- Check that synthetic watermark does not interfere with listening exercises.
Get Feedback
Share generated audio clips with colleagues or student focus groups to gauge engagement and pronunciation accuracy. Since the API does not offer custom voice cloning or emotion tags, rely on the inherent tones of voices like Luna for storytelling or Sage for instructions to determine what resonates best with your specific audience.
- Collect student reactions to different voice personalities.
- Confirm pronunciation clarity for non-native speakers.
- Validate that the tone matches the lesson context.
Final Recommendations
Your French language learning API integration is now ready to support diverse classroom needs. By combining these eight distinct voices, you can create engaging, multi-modal listening materials that keep students interested and attentive throughout the semester.
Mix and Match
Diversify your audio content by alternating between instructional voices like Sage and narrative ones like Luna. This variety prevents auditory fatigue during long reading sessions and helps students distinguish between different speech styles, from formal news delivery to casual conversation.
- Use Aria for clear pronunciation drills.
- Switch to Milo for interactive dialogues.
- Apply Felix for energetic announcements.
Consistency for Learners
Maintain consistent voice assignments for specific recurring characters or lesson types. When students hear the same voice for a particular story or grammar rule, they build stronger cognitive associations, making it easier to focus on vocabulary and syntax rather than adapting to new tones.
- Assign one voice per story series.
- Keep instructional voices stable.
- Use neutral tones for factual content.
Regular Updates
Refresh your audio library periodically to reflect new vocabulary or current events. With a $0.01 per 1,000 characters cost, generating new materials is affordable. Use the free trial credit to test new voice combinations before committing to larger batches.
- Update seasonal reading materials.
- Test new voice pairs.
- Monitor character usage limits.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Selecting the right voice is essential for maintaining student engagement during language learning.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- LibriTTS-R corpus (OpenSLR 141)
- MDN: Web Speech API
- OpenAI API reference: create speech
- openai-node on GitHub
- Node.js documentation
- W3C WAI: making audio and video accessible
- Python documentation
- ISO 639 language codes
- LibriTTS-R paper (arXiv 2305.18802)
- Creative Commons Attribution 4.0
- TRON developer hub: TRC-20
