Frequently asked questions
What is the best voice for a museum tour?
Q: Which voice is best for a museum tour? A: There is no single best voice; it depends on your content. For calm, instructional content, use Sage. For a warm, clear narrator style, choose Aria. Luna offers a soft storytelling tone, while Atlas provides a neutral news style. Select the voice that matches the specific mood of your exhibit.
Can I mix male and female voices in one guide?
Yes, you can mix voices by splitting your text into separate requests. Each request supports one of eight presets. You can combine the resulting audio files to create a guide with multiple narrators, such as alternating between Felix for energetic segments and Luna for storytelling.
Does the voice sound natural in multiple languages?
Yes, any voice can speak any of the 23 supported languages. For non-English Latin-script text, you must set the language parameter. If omitted, Latin text defaults to English. The API supports Arabic, Chinese, French, German, Japanese, and others without requiring different voice models.
What is the difference between tts-1 and tts-1-hd?
Both models use the same engine and price. The only difference is the model identifier. You can use either "tts-1" or "tts-1-hd" in your requests. They produce identical audio quality and cost the same $10 per 1 million characters, offering no performance distinction.
How do I make the voice sound slower?
Adjust the speed parameter between 0.25 and 4.0. The default is 1.0. Set it lower, such as 0.8, to slow the narration. This is useful for complex museum exhibits where visitors need more time to read signs. The API processes these requests sequentially.
Can I use the same voice for my hotel app?
Yes, the API is designed for scalable audio content. You can use any of the eight voices for hotel apps, audioguides, or video voiceovers. There are no restrictions on industry use. Each request costs $0.01 per 1,000 characters, making it cost-effective for large apps.
Are the voices based on real people?
Yes, the voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. Each generated file carries an inaudible watermark marking it as synthetic speech. The voices are presets, not custom clones of specific individuals.
Can I clone my own voice for the museum?
No, custom voice cloning is not offered. You cannot upload your own voice or design a new one. You must choose from the eight available presets: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, or Orion. These are the only voices available for generation.
Do the voices sound robotic?
The voices are based on licensed human recordings and include an inaudible synthetic watermark. They do not use SSML or emotion tags, but they are not robotic. They offer clear, natural narration suitable for tourism and museum contexts without hiring voice actors.
Can I change the voice after generating audio?
No, the audio file is generated based on the voice parameter in the request. Once generated, the audio is returned and not stored by the API. To change the voice, you must send a new request with the desired voice preset and input text.
What happens if I change my mind on voice?
You can simply send a new request with a different voice preset. Failed requests are free, so changing your mind does not incur extra costs if the generation fails. Otherwise, you pay per character generated. There are no monthly plans or subscriptions.
Is there a voice for children's tours?
There is no specific children's voice, but Luna offers a soft, storytelling tone suitable for younger audiences. Nova is bright and upbeat. You can adjust the speed to be slower for clarity. Choose the preset that best matches the energy of your content.
Can I use these voices for YouTube videos?
Yes, the API is suitable for video voiceovers and scalable audio content. You can generate audio files in MP3, WAV, OPUS, FLAC, or PCM formats. The voices are based on licensed recordings, and each file carries an inaudible synthetic watermark. The API does not require attribution, but the underlying dataset (LibriTTS-R) is CC BY 4.0, which may require attribution depending on your use case.
Do I need to attribute the voice in my app?
The API does not require specific attribution for the voice preset. The voices are based on licensed recordings from LibriVox, but the generated audio is a synthetic product. You do not need to credit the API or the voice actors in your app interface.
Understanding Voice Personalities in Tourism
Museum audioguides require distinct tonal qualities to match exhibit themes. Selecting the right voice personality ensures historical narratives feel authentic rather than generic. Our API offers eight preset voices, each designed for specific communication styles, allowing curators to align audio with visitor expectations.
Warm and Clear: Aria for Storying
Aria delivers a warm, clear narration style ideal for general storytelling. Her tone guides visitors through exhibits with clarity, making complex historical contexts accessible. This voice works well for introductory panels where clarity is paramount.
- Aria is described as a warm, clear narrator.
Bright and Upbeat: Nova for Youth Tours
Nova offers a bright, upbeat delivery suitable for younger demographics or interactive tours. Her energetic tone keeps attention high during fast-paced exhibits or digital scavenger hunts. This preset is effective for engaging visitors who prefer a lively experience.
- Nova is described as bright and upbeat.
Soft and Storytelling: Luna for History
Luna provides a soft, storytelling voice perfect for historical narratives. Her gentle tone suits reflective areas or poignant historical moments. Using Luna helps create an immersive atmosphere without overpowering the visual content of the exhibit.
- Luna is described as soft and storytelling.
Male Narrator Options for Professional Tone
For museums and tourism startups, male voices offer distinct professional tones that align with specific content types. Whether promoting a new exhibit or guiding a casual tour, selecting the right voice ensures clarity and engagement without requiring multiple voice actors.
Energetic Felix for Promos
Felix delivers an energetic, promo-style narration ideal for highlighting special exhibits or limited-time offers. His tone captures attention quickly, making him suitable for short promotional clips or social media teasers.
- Best for: Promotional videos and exhibit highlights
Friendly Milo for Conversational Tours
Milo provides a friendly, conversational tone perfect for standard audio guides. His approachable delivery helps visitors feel guided rather than lectured, enhancing the experience for casual tourists.
- Best for: General museum tours and self-guided walks
Neutral Atlas for News-Style Updates
Atlas offers a neutral, news-style voice suitable for factual updates or historical information. His clear, unbiased delivery ensures that information is perceived as authoritative and reliable.
- Best for: Historical facts, news-style updates, and informational signage
Matching Voice to Museum Theme
Selecting the right narrator voice depends heavily on your museum’s specific theme and the type of experience you want to create for your visitors.
Calming Sage for Instructional Content
Sage offers a calm, instructional tone ideal for wayfinding or rule explanations. Its neutral delivery ensures visitors absorb practical information without distraction, making it perfect for entrance guidelines or safety instructions.
Deep Orion for Documentary Style
Onion provides a deep, documentary-style narration suitable for historical exhibits or deep-dive content. This voice adds gravitas to complex narratives, enhancing the educational value of detailed exhibit descriptions.
Consistency Across Multiple Languages
Any voice can speak any of the 23 supported languages, including English, Spanish, and German. This allows you to maintain a consistent brand voice across different language versions of your audioguide, ensuring a unified visitor experience regardless of the language setting.
Technical Considerations for Voice Selection
Selecting the right technical parameters ensures your museum audioguide sounds professional and remains accessible across different devices. The API allows precise control over delivery speed and output format, letting you balance clarity with file size without needing complex post-production tools.
Language Compatibility with Voices
All eight voices support the full list of 23 languages, including Arabic, Chinese, and Swahili. This means you can maintain a consistent narrator personality across all your museum branches, regardless of the local language spoken by visitors. If the language is not specified, the system guesses from the script, but setting it explicitly ensures accurate pronunciation for non-English Latin texts.
- Supports 23 languages with all 8 voice presets.
- Explicit language codes prevent mispronunciation in mixed-script texts.
- Consistent voice identity across global museum locations.
Speed Adjustment for Clarity
You can adjust speech speed from 0.25 to 4.0. For detailed exhibit descriptions, slowing the speed to 0.75–0.9 improves comprehension for non-native speakers and elderly visitors. Conversely, faster speeds work well for promotional snippets or short navigation instructions where brevity is key.
- Speed range: 0.25 to 4.0.
- Slower speeds enhance clarity for complex historical facts.
- Adjust pacing based on visitor demographics and content density.
Audio Format Impact on Perception
Choose MP3 for efficient storage, or WAV/FLAC for high-fidelity museum displays. MP3 files are smaller, while uncompressed formats preserve the full nuance of the voice for high-end audio installations. PCM is available for raw processing needs.
- MP3: Best for mobile apps and quick downloads.
- WAV/FLAC: Ideal for high-quality offline museum kiosks.
- PCM: For raw audio integration in custom software.
Audiencing and Demographics
Selecting a narrator voice requires aligning audio characteristics with your target demographic’s listening habits and cognitive load. Young tourists often prefer the bright, upbeat energy of Nova for interactive mobile tours, while seniors may find the calm, instructional tone of Sage easier to follow in noisy gallery environments. This distinction ensures clarity without straining the listener’s attention.
Demographic considerations also extend to pacing and volume. Younger audiences might tolerate faster speeds (up to 1.5x), whereas older visitors often benefit from slower, clearer articulation found in Sage or Luna. Testing these variations against your specific exhibit layout helps optimize retention and engagement across different visitor segments.
Voice Preferences for Young Tourists
Younger demographics, particularly those using audio guides on smartphones, often respond well to higher energy and dynamic pacing. Voices like Nova or Felix provide a modern, engaging feel that matches the visual stimulation of contemporary exhibits. Their upbeat nature keeps attention levels high during fast-paced tours or interactive museum games.
For this group, speed settings can be pushed higher (1.5–2.0x) without losing comprehension, allowing for shorter, more efficient tour experiences. The clear, crisp production quality of the TTS engine ensures that audio remains distinct even in crowded, noisy gallery spaces where background chatter is common.
- Nova suits interactive, fast-paced youth tours
- Felix works well for promotional video voiceovers
- Higher speed settings (1.5x–2.0x) are effective for mobile users
Senior-Friendly Voice Characteristics
Older visitors often prioritize clarity and comfort over energy. The soft, storytelling tone of Luna or the calm, instructional voice of Sage reduces cognitive load, making complex historical or scientific explanations easier to digest. These voices offer a steady, predictable rhythm that aids in processing information without the fatigue associated with faster, more energetic narrations.
Slower speech rates (0.8–1.0x) combined with these warmer tones enhance comprehension for seniors. The consistent, neutral delivery avoids sudden pitch changes or high-energy bursts that might startle or distract visitors who are focusing heavily on visual details or reading accompanying plaques.
- Sage provides calm, instructional clarity
- Luna offers soft, storytelling comfort
- Slower speeds (0.8–1.0x) improve comprehension
Cultural Nuances in Voice Selection
While all eight voices can speak all 23 supported languages, the perceived 'tone' remains tied to the voice’s inherent character. For instance, Atlas’s neutral, news-style tone works well for factual, historical updates in any language, providing a consistent brand voice across international branches. Conversely, Orion’s deep, documentary style adds gravitas suitable for solemn historical sites.
When localizing content, ensure the language parameter is explicitly set, especially for Latin-script languages other than English, to avoid automatic English fallback. This ensures that the cultural nuance of the narration matches the local audience’s expectations, whether it’s a friendly conversational tone in Spanish or a precise instructional style in German.
- Atlas ensures neutral consistency across languages
- Orion adds gravitas to historical content
- Explicit
languageparameter prevents fallback errors
Testing Voices Before Production
Before committing to full-scale production, validate your voice selection using the $0.50 free trial. This credit covers approximately 50,000 characters, allowing you to test all eight preset voices—Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion—without financial risk. Generate short script segments for each voice to evaluate their suitability for your specific exhibit context.
Use the trial to verify technical compatibility, ensuring the chosen voice renders correctly across your target languages and formats. This step confirms that the synthetic watermark and audio quality meet your standards before you begin the character-count billing cycle.
Using the Free Trial for Testing
The $0.50 trial credit provides about 50,000 characters, valid for 7 days. Use this to generate samples for all eight voices: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion. Test scripts in multiple languages to ensure the voice matches the intended tone, whether warm narration or news-style updates.
Create short segments (under 4,000 characters) for each voice to compare pacing and clarity. Since requests are processed sequentially, test during low-load periods to avoid queue delays. This allows you to select the single best voice before scaling up to full guide production.
- Generate samples for all 8 voices
- Test in target languages
- Verify character limits (4,096 chars/max)
Comparing MP3 vs WAV Quality
MP3 offers smaller file sizes, ideal for mobile apps and web delivery, while WAV provides uncompressed quality for high-fidelity in-museum speakers. Both formats are supported, but MP3 is generally sufficient for clear voiceovers. Test both formats on your actual hardware to hear the difference.
Note that audio files are not stored server-side; you receive them directly in the response. Download and compare samples locally to decide which format best preserves the voice's nuance without excessive bandwidth usage for your users.
- MP3 for mobile/web efficiency
- WAV for high-fidelity speakers
- No server storage of audio
Gathering Feedback from Focus Groups
Share generated samples with small groups of target visitors, such as young tourists or seniors, to gauge preference. Ask specific questions about clarity, warmth, and engagement. Since the API returns synthetic speech with an inaudible watermark, focus on the auditory experience rather than identifying it as AI.
Collect feedback on pacing and volume. Adjust the speed parameter (0.25–4.0) based on listener input. Slower speeds may improve comprehension for elderly visitors, while faster speeds might suit quick audio tours. Use this data to finalize your voice choice before bulk generation.
- Test with target demographics
- Adjust speed (0.25–4.0)
- Focus on clarity and warmth
Licensing and Usage Rights
The voices in our TTS API are derived from licensed recordings of the LibriTTS-R dataset, which originates from LibriVox public-domain audiobook readings. This foundation ensures that the synthetic audio is built upon legally cleared source material, providing a stable licensing baseline for museum and tourism applications.
LibriVox Public Domain Source
LibriVox volunteers recorded books that have entered the public domain in the United States. Our engine samples these recordings to create the eight voice presets, ensuring that the underlying vocal characteristics are free from restrictive copyright claims.
- Source: LibriVox public-domain audiobooks
- Dataset: LibriTTS-R
CC BY 4.0 License Implications
The LibriTTS-R dataset is licensed under CC BY 4.0. While this allows commercial use, it may require attribution. You should verify if your specific museum or tour operator application requires explicit attribution to the LibriVox project or the original readers in your audio guides or website metadata.
- License: CC BY 4.0
- Requirement: Attribution may be needed
Commercial Use Permissions
The CC BY 4.0 license permits commercial use, making these voices suitable for paid audioguides, video voiceovers, and ticketed museum tours. There are no per-use fees for the voice license itself, separate from the API character-based pricing. You can generate unlimited audio for commercial distribution without additional royalty payments.
- Commercial use allowed
- No royalty payments

Common Mistakes in Voice Selection
Selecting the wrong narrator undermines the educational value of your audio guide. Curators often pick voices based on personal preference rather than how the content will be consumed in a noisy, dynamic environment.
Choosing Too Many Voices
Using more than two distinct voices in a single exhibit confuses visitors and dilutes brand identity. Consistency builds trust; switching between Aria and Felix mid-tour feels disjointed.
Ignoring Background Noise Compatibility
Museum halls often have echo or ambient noise. A deep voice like Orion may become muddy in reverberant spaces, while a bright voice like Nova cuts through better. Test audio in actual locations before finalizing.
Mismatching Tone with Content
Using Felix for a solemn memorial creates tonal dissonance. Match the voice’s inherent character to the exhibit’s mood: use Luna for storytelling history, but avoid her for urgent safety instructions where Atlas’s neutrality is clearer.
Voice Consistency Across Locations
Maintaining a unified auditory identity across multiple museum branches ensures visitors recognize your institution regardless of location. Using a single voice preset for all branch offices eliminates the jarring effect of switching narrators between locations.
Standardizing tone across regions prevents brand dilution, especially when content is localized into different languages. Since any voice can speak all 23 supported languages, you can deploy the same narrator personality globally without sacrificing linguistic authenticity.
Using the Same Voice for Branch Offices
Deploying a consistent voice across all branches strengthens brand recall. Visitors transitioning from a main site to a branch experience a familiar auditory cue, reinforcing trust in the institution’s narrative authority.
- Use the same voice preset for all branch locations to maintain continuity.
- Ensure the voice works well in multiple languages if branches serve diverse communities.
Standardizing Tone Across Regions
Standardization simplifies content management. Instead of managing different voice contracts or recordings for each region, you manage one audio engine configuration. This reduces complexity and ensures consistent quality control across all outputs.
Whether the content is a historical tour in English or a promotional clip in Spanish, the underlying vocal character remains identical, preserving the institution's core identity.
- Centralize voice selection in your API configuration.
- Apply the same speed and format settings across all regions for uniform playback experience.
Brand Voice Guidelines
Define clear guidelines for which voice matches your brand’s personality. For example, use Aria for warm storytelling or Atlas for neutral updates. Document these choices to ensure new content creators select the correct preset.
Include voice selection criteria in your content production workflow. This ensures that even as you scale to new languages or regions, the auditory brand remains distinct and professional.
- Document preferred voices for different content types (e.g., tours vs. news).
- Review voice performance regularly to ensure it aligns with evolving brand standards.
Accessibility and Voice Choice
To ensure accessibility for visually impaired visitors, clarity and articulation are paramount. Our TTS engine delivers natural-sounding narration that avoids the robotic cadence often found in older systems, allowing users to focus on exhibit details without distraction.
Proper pacing ensures every word is distinct, supporting cognitive load during audio guide tours. By selecting voices like Aria or Luna, you provide a warm, clear listening experience that enhances engagement and comprehension for all users.
Clarity for Visually Impaired Visitors
High-fidelity audio output ensures that historical narratives are delivered clearly. The synthetic watermark remains inaudible, preserving the immersive experience without drawing attention to the technology itself.
- Natural intonation reduces listener fatigue during long tours
- Clear enunciation aids comprehension for non-native speakers
Pacing and Articulation Standards
Adjustable speed (0.25–4.0x) allows visitors to control playback, accommodating individual processing speeds. This flexibility is critical for elderly visitors or those with hearing impairments who may need slower narration to capture details.
Consistent articulation across 23 languages ensures that multilingual guides maintain high quality, regardless of the visitor's preferred language. This consistency supports a seamless experience across diverse tourist demographics.
- Speed control from 0.25x to 4.0x
- Consistent articulation across 23 languages
Voice Diversity in Exhibits
Using different voices for different sections can signal topic shifts, aiding navigation through complex exhibits. For instance, a deep, documentary-style voice like Orion suits historical timelines, while a bright tone like Nova fits interactive youth zones.
Diversity in voice selection reflects the variety of content, enhancing the overall museum narrative. This approach keeps visitors engaged and provides auditory cues that help structure the tour experience.
- Use Orion for historical timelines
- Use Nova for interactive youth zones
Future-Proofing Your Audio Content
As your collection grows, the API’s character limit of 4,096 allows you to split long exhibits into manageable chunks while maintaining a consistent narrator identity. You can generate audio for hundreds of locations without re-hiring voice actors, ensuring brand coherence across new branches or digital updates.
Scalability of Voice Selection
The eight preset voices are language-agnostic, meaning you can deploy the same narrator across all 23 supported languages without losing tonal consistency. This eliminates the need for separate casting sessions as you expand into new regions or add new exhibition halls.
- Generate audio for unlimited exhibits using the same voice preset.
- Scale from 10 to 10,000 audio files without additional casting costs.
- Maintain consistent branding across all physical and digital touchpoints.
Updating Narrations Without Recasting
When exhibit content changes, you simply regenerate the specific text segment. Since input is not stored for training, you can update facts or descriptions instantly without worrying about voice drift or licensing complications. This ensures your audioguide always reflects the current state of the museum.
Digital vs Physical Tour Guides
The same API endpoint serves both in-app tours and downloadable MP3/WAV files for physical devices. By standardizing on one voice profile, you ensure that visitors receive the same high-quality experience whether they are using a museum’s mobile app or renting a standalone audio guide unit.
Support and Feedback Channels
Our support channel is email-only. There is no phone number or live chat, ensuring every query is documented and addressed systematically. This approach suits museum curators who prefer asynchronous communication that fits around exhibit hours.
When submitting a ticket, include your API key, the specific voice used, and the request timestamp. Our team reviews audio samples to verify synthesis quality against our standards.
Contacting Support via Email
Send your inquiries to the address provided on the /contact/ page. Include your account key for faster identification. We aim to respond within one business day, though complex audio artifacts may require additional time for analysis.
Provide the exact text submitted and the generated audio file if possible. This helps us reproduce the issue precisely without asking for redundant information.
- Include your API key
- Attach the generated audio file
- Specify the voice model used
Reporting Voice Quality Issues
If you notice artifacts, unnatural pacing, or incorrect language detection, describe the specific segment. Mention the ISO language code you set versus what was generated. Our engineers check the server logs to determine if it was a processing error or a model limitation.
Failed requests are free, so if the audio is unusable, you can retry or contact us for a refund confirmation. We do not guarantee perfect pronunciation for every rare proper noun in all 23 supported languages.
- Describe the artifact type
- Specify the input text
- Note the language code used
Requesting New Voice Features
We currently offer eight fixed presets based on LibriVox recordings. We do not offer custom voice cloning or SSML support, so requests for these features will be noted but not prioritized for immediate implementation. Our roadmap focuses on stability and new language support rather than new voice models.
You can suggest improvements for existing voices, such as speed adjustments or format preferences. While we cannot add emotion tags, we may adjust the default speed profiles for specific voices based on user feedback from tourism partners.
- Suggest voice adjustments
- Request new languages
- Note missing features
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This clarity helps developers quickly prototype audio guides for museums or tourist apps without guessing parameter formats or encoding requirements.
23 languages
Any voice speaks any of these languages.
