Frequently asked questions
What is the best voice for news?
Atlas is designed as a neutral, news-style voice. It provides a clear and professional tone suitable for public information systems. While other voices can read news, Atlas offers a consistent, neutral delivery that works well for standard reporting without adding unnecessary emotional inflection.
How do I detect language automatically?
The API guesses the language from the script if you do not set the language field. Latin script text defaults to English. For non-English Latin-script text, you should explicitly set the language parameter to ensure accurate synthesis and avoid mispronunciation.
Is Aria suitable for storytelling?
Aria is described as a warm, clear narrator. While Luna is specifically noted for soft storytelling, Aria’s warm tone makes it suitable for general narration. It provides a clear, engaging delivery that works well for reading texts to visually impaired users.
Can Atlas be used for emergency alerts?
Atlas offers a neutral, news-style voice that is clear and easy to understand. Its neutral tone ensures the message is delivered without distracting emotional inflection, making it a reliable choice for emergency alerts where clarity and directness are prioritized over expressiveness.
Does Luna sound too soft for instructions?
Luna is described as soft and suited for storytelling. For instructional content, Sage is recommended as it is calm and instructional. Sage provides a clearer, more directive tone better suited for guiding users, whereas Luna’s softness might be less distinct for commands.
Is Milo friendly enough for children?
Milo is described as friendly and conversational. This tone makes it suitable for engaging audiences or creating a welcoming atmosphere in public information systems. Its conversational style helps make text accessible and less formal for non-native speakers or children.
How does Orion differ from Atlas?
Atlas is a neutral, news-style voice, while Orion is a deep, documentary-style voice. Atlas provides a standard, clear delivery for general information, whereas Orion offers a deeper, more authoritative tone suitable for documentary-style content or more serious, immersive public information projects.
Can I mix voices in one project?
The API processes one voice per request. To mix voices, you must send separate requests for each segment of text using the desired voice. You then join the resulting audio files. There is no single endpoint to generate multiple voices in one file.
Do voices sound different in other languages?
Any voice can speak any of the 23 supported languages. The vocal characteristics remain consistent across languages. However, pronunciation and intonation will adapt to the specific language spoken, ensuring natural-sounding speech in Arabic, Chinese, English, Spanish, and other supported languages.
Is the watermark audible?
Every generated audio file carries an inaudible watermark that marks it as synthetic speech. This watermark is not audible to listeners. It serves as a marker to identify the audio as AI-generated without affecting the user experience or audio quality.
Can I use these voices commercially?
The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings. You can use the generated audio commercially. The API provides a straightforward service without enterprise contracts, allowing you to integrate these voices into your public information systems.
What is the speed range?
The speed parameter accepts values between 0.25 and 4.0. This allows you to adjust the playback speed of the generated audio. Slower speeds may improve clarity for visually impaired users, while faster speeds can reduce generation time for long texts.
Do I need to specify speed?
No, specifying speed is optional. If you do not set the speed parameter, the API uses a default speed of 1.0. You can adjust it between 0.25 and 4.0 if you need slower or faster audio for specific accessibility or content needs.
Are there emotion tags available?
No, emotion tags are not available. The API does not support SSML, emotion tags, or phoneme dictionaries. You control tone primarily through voice selection and speed. For example, use Milo for a friendly tone or Atlas for a neutral news style.
Understanding voice tones
Public information systems require audio that matches the content’s intent. Neutral voices like Atlas deliver news with clarity, while warm tones like Aria suit community stories. Choosing the right personality ensures users perceive the message correctly, whether it is a urgent alert or a gentle reminder.
Neutral vs. warm personalities
Neutral voices such as Atlas provide a straightforward, news-like delivery ideal for public notices. Warm voices like Aria offer a clear, narrative tone suitable for community updates or longer informational segments. Selecting between these depends on whether your goal is rapid information transfer or engagement.
- Atlas: Neutral, news-style delivery for urgent updates
- Aria: Warm, clear narrator for community stories
- Milo: Friendly, conversational tone for general inquiries
Instructional vs. storytelling
Instructional content benefits from Sage’s calm, structured voice, which aids comprehension for non-native speakers. Storytelling or cultural content suits Luna’s soft, narrative style. Using the correct voice type reduces cognitive load for visually impaired users listening to accessibility tools.
- Sage: Calm, instructional voice for step-by-step guides
- Luna: Soft, storytelling tone for cultural content
Matching tone to content type
Align voice selection with content purpose. Use Felix for energetic promotional announcements or Milo for conversational service updates. Consistent tone across your system builds user trust and ensures accessibility standards are met without requiring complex configuration.
- Felix: Energetic, promo-style for announcements
- Milo: Friendly, conversational for service updates
- Orion: Deep, documentary style for detailed reports
Language and dialect considerations
Selecting the right voice requires precise language handling to ensure public information is accessible to non-native speakers and diverse communities. Our API supports 23 languages, which allows indie developers to serve multilingual audiences without complex backend logic.
Support for 23 languages
We support Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. Any voice can speak any of these languages, providing flexibility for public service announcements.
- Arabic, Chinese, Danish, Dutch, English
- Finnish, French, German, Greek, Hebrew, Hindi
- Italian, Japanese, Korean, Malay, Norwegian, Polish
- Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Automatic language detection
The API guesses the language from the script if the language field is omitted. Latin script text defaults to English, which works well for standard English content but may cause errors for other Latin-based languages like French or Spanish without explicit coding.
Manual override for accuracy
For non-English Latin-script text, you must set the language parameter to the ISO code. This ensures correct pronunciation and prevents the system from defaulting to English, which is critical for accurate public notices and accessibility tools.
- Set
languagefor non-English Latin text - Use ISO codes for precision
- Avoids default English misinterpretation
Unsupported languages
Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali and all other languages not in the list are not supported. Ensure your target audience’s language is in the list before integrating, as these are common public information languages that may require alternative solutions.
Accessibility standards
Accessibility is critical for public information systems. Our API delivers clear, human-like audio that ensures content is accessible to visually impaired users and non-native speakers without requiring complex enterprise contracts or specialized hardware.
Consistency in tone and pacing builds trust in public notices. By standardizing voice output, you ensure that critical information is delivered uniformly across all platforms, from websites to digital kiosks.
Clarity for visually impaired users
The synthetic voices are designed for natural intelligibility, reducing listener fatigue during long readings. Each voice offers a distinct tone that helps users distinguish between different types of content, such as news updates versus instructional guides.
The engine produces natural-sounding speech, making it easier for assistive technology users to follow complex sentences without losing context.
- Natural intonation reduces cognitive load
- Distinct voice personalities aid content differentiation
Consistency in public notices
Public information systems require reliable audio output. Using a fixed voice preset ensures that your audience recognizes the source of information immediately, whether it’s a weather alert or a community event.
This consistency holds true across all 23 supported languages, allowing NGOs and local governments to maintain a unified brand voice even when serving diverse linguistic communities.
- Unified audio identity across regions
- Reliable output for emergency alerts
Speed and pacing options
Users control their listening experience with the speed parameter, adjustable from 0.25 to 4.0. This allows visually impaired users to slow down complex texts or speed through simple updates.
For public notices, a moderate speed often provides a good balance between information density and comprehension, ensuring critical details are not missed.
- Adjustable speed (0.25–4.0x)
- Optimal pacing for complex data
Technical implementation
Integrating voice selection requires sending a POST request to the speech endpoint with the chosen voice preset. The API accepts standard JSON bodies, allowing you to specify the model, input text, and voice identifier directly in your code.
Integrating voice selection in code
Use the OpenAI SDK by setting the base URL to https://api.b777x.info/v1. Pass your Bearer token and select one of the eight presets like 'aria' or 'atlas' in the voice field. This ensures the audio matches your accessibility or public information needs.
- Set base_url to the API domain
- Include Authorization Bearer token
- Pass voice ID in JSON body
Handling voice parameters
Specify the model as 'tts-1' or 'tts-1-hd'. You can adjust speed between 0.25 and 4.0 to suit different reading speeds for visually impaired users. The language parameter is optional but recommended for non-English Latin text to ensure accurate pronunciation.
- Model: tts-1 or tts-1-hd
- Speed range: 0.25–4.0
- Language: optional ISO code
Testing audio output quality
Generate short samples using the free trial credit. Listen for clarity and tone consistency across different voices. Check that the audio format (mp3, wav, etc.) meets your player requirements. Verify that the synthetic watermark is present but inaudible.
- Use free trial for samples
- Check clarity and tone
- Verify format compatibility
Cost vs. quality trade-offs
The core pricing remains constant at $10 per million characters regardless of model choice. Both "tts-1" and "tts-1-hd" use the same engine and incur identical charges, allowing you to prioritize audio fidelity without budget surprises.
tts-1 vs. tts-1-hd models
Select "tts-1-hd" only when higher acoustic quality justifies the processing time. Since both models share the same rate and character limits, the decision rests solely on your accessibility requirements rather than cost.
- Both models cost $0.01 per 1,000 characters.
- No price difference between standard and high-definition.
Format selection impact
Output formats like mp3, wav, opus, flac, or pcm do not affect the character count or the final invoice amount. Choose the format that best fits your player’s decoding capabilities without worrying about variable billing.
Budgeting for high-volume use
With a minimum top-up of $5 via USDT on TRON, small NGOs can start immediately. Since failed requests are free, your billing reflects only successful audio generation, making cost prediction straightforward based on character volume.
- Minimum top-up: $5 USDT (TRC-20).
- Failed requests are refunded in full.
Voice presets overview
Selecting the right voice ensures your public information is both accessible and engaging. Our eight presets cover a spectrum from warm narration to neutral reporting, allowing you to match the tone of your text-to-speech output to specific content types.
Female presets: Aria, Nova, Luna, Sage
Aria provides a warm, clear narration suitable for general reading. Nova offers a bright, upbeat tone for announcements. Luna delivers a soft, storytelling style for longer texts. Sage is calm and instructional, ideal for guides.
Male presets: Felix, Milo, Atlas, Orion
Felix is energetic and promo-focused. Milo sounds friendly and conversational. Atlas provides a neutral, news-like delivery. Orion offers a deep, documentary style for serious content.
Use cases for each voice
Use Atlas for consistent public notices and Nova for urgent updates. Sage works well for instructional accessibility tools, while Luna suits narrative content. Felix and Orion enhance promotional or documentary-style audio.
- Aria: General narration
- Nova: Upbeat announcements
- Luna: Storytelling
- Sage: Instructional
- Felix: Energetic promo
- Milo: Conversational
- Atlas: Neutral news
- Orion: Deep documentary
Community feedback
Independent developers building accessibility tools for NGOs often prioritize clarity over stylistic flair. Because the API processes requests sequentially, consistent voice selection helps maintain user familiarity across long public information broadcasts, even when queue times vary during peak hours.
User preferences in NGOs
NGO developers frequently select Sage for instructional content or Luna for storytelling segments. The neutral tone of Atlas is preferred for news-style updates where distraction must be minimized. These choices support non-native speakers by providing predictable speech patterns.
- Sage for clear instructions
- Luna for narrative segments
- Atlas for neutral news updates
Public sector adoption rates
Adoption is driven by the straightforward integration via standard HTTP requests. Teams using the OpenAI SDK with a custom base_url report smooth deployment. The lack of SSML requirements reduces development time for public notice systems.
Common voice selections
Aria and Nova are common for general public announcements due to their clarity. Milo is often chosen for conversational-style updates. Developers test these presets against their specific text lengths to ensure the 4096-character limit fits their needs.

Licensing and attribution
Our voice presets are based on licensed recordings from LibriTTS-R, derived from LibriVox public-domain audiobook readings. This foundation ensures that the acoustic quality meets professional standards while remaining accessible for public information systems.
LibriVox public domain source
The underlying voice models utilize high-quality recordings from the LibriVox project. These sources are freely available and form the basis for the eight distinct voice presets offered in the API.
CC BY 4.0 compliance
We adhere to the CC BY 4.0 license requirements. By using these voices, you are integrating content that respects the original creators' contributions while allowing flexible usage in your accessibility tools.
Synthetic watermark explanation
Every generated audio file carries an inaudible watermark marking it as synthetic speech. This transparently identifies the audio as AI-generated without affecting the listening experience for end users.
Testing your selection
Before committing to a production voice, use the $0.50 free trial to test synthesis quality. This credit covers approximately 50,000 characters, allowing you to validate audio output without immediate payment. Since the trial expires in 7 days, generate samples early to ensure the selected voice meets your accessibility standards.
You can test all eight presets and various formats to determine which combination best suits your specific content type. This approach helps you avoid costly mistakes when scaling your public information system.
Using the free trial
Sign up with Google to receive $0.50 in credit. This amount allows for roughly 50,000 characters of generation, which is sufficient to test multiple voices and speeds. Note that email-based accounts do not receive this trial credit, so plan your testing accordingly.
Use the trial to verify that the voice clarity meets your needs for visually impaired users or non-native speakers. You can adjust the speed parameter between 0.25 and 4.0 to find the optimal pacing for long public notices.
- Google sign-up grants $0.50 trial credit.
- Credit is valid for 7 days.
- Email-based accounts start with zero balance.
- Test multiple voices before purchasing USDT.
Generating sample audio
Send POST requests to the speech endpoint with short text samples. Test different response formats like mp3, wav, opus, or flac to see which best fits your player. Remember that each request is limited to 4096 characters, so split longer texts if needed.
Use the official OpenAI SDK by changing the base_url to https://api.b777x.info/v1. This allows you to reuse familiar code while accessing our voices. Generate samples for each voice preset to compare tone and clarity side-by-side.
- Use POST /v1/audio/speech for samples.
- Test mp3, wav, opus, and flac formats.
- Limit input to 4096 characters per request.
- Use OpenAI SDK with adjusted base_url.
Evaluating clarity and tone
Listen for natural pacing and clear articulation, especially for public information where comprehension is key. Check if the voice sounds too robotic or overly emotional for your context. For example, use Atlas for neutral news or Luna for storytelling.
Verify that the synthetic watermark is present but not intrusive. Ensure that the audio quality remains consistent across different speeds and languages. This evaluation helps you choose the right voice for your specific audience before scaling up.
- Check for natural pacing and clarity.
- Ensure synthetic watermark is present.
- Test consistency across speeds and languages.
- Compare tone against content type.
Common mistakes to avoid
Selecting the wrong voice or misconfiguring language settings can undermine the clarity of public notices. By avoiding these common pitfalls, you ensure that your accessibility tools remain effective and understandable for diverse audiences.
These errors are easy to prevent with a quick review of your API parameters before deployment.
Ignoring language detection
Latin-script text defaults to English if the language field is omitted. For non-English Latin text, you must explicitly set the language parameter to avoid incorrect pronunciation.
If your audience speaks a language not in the supported list, such as Vietnamese or Ukrainian, the system will still generate audio, but it may sound unnatural.
- Explicitly set the
languageISO code for non-English Latin text - Verify target languages against the supported list of 23 languages
Choosing wrong tone for context
Using 'felix' for a solemn public notice or 'aria' for a fast-paced promo can create a mismatch between audio and visual context.
Match the voice personality to the content: 'atlas' for news, 'sage' for instructions, 'luna' for storytelling.
Consistency in tone builds trust with visually impaired users who rely on audio cues.
- Use 'atlas' for neutral news updates
- Use 'sage' for instructional content
- Avoid 'nova' for serious announcements
Overlooking character limits
Each request is limited to 4096 characters. Exceeding this limit will cause the request to fail.
Split longer texts into parts and join the audio files manually or via your client library.
Remember that spaces and punctuation count toward the character limit.
- Limit input text to 4096 characters per request
- Split long texts into manageable chunks
- Count spaces and punctuation as characters
Future-proofing your API
Scaling your public information system is straightforward with our usage-based pricing. There are no monthly subscriptions or minimum spend requirements, allowing you to adjust costs directly to your output volume. This model suits indie developers who may experience fluctuating demand for accessibility audio or public notices.
Funding your account is simple via USDT (TRC-20) on the TRON network. Top-ups are credited automatically within minutes, ensuring your service remains active without administrative delays or complex invoicing processes.
Scalability with USDT payments
Scaling your public information system is straightforward with our usage-based pricing. There are no monthly subscriptions or minimum spend requirements, allowing you to adjust costs directly to your output volume. This model suits indie developers who may experience fluctuating demand for accessibility audio or public notices.
Funding your account is simple via USDT (TRC-20) on the TRON network. Top-ups are credited automatically within minutes, ensuring your service remains active without administrative delays or complex invoicing processes.
Updating voice selections
You can change your voice configuration at any time by updating the voice parameter in your API requests. Whether you need a calm instructional tone for technical guides or a warm narrator for community updates, the change takes effect immediately with your next call.
This flexibility allows you to A/B test different voices for clarity or audience engagement without needing to reconfigure your entire integration pipeline.
- Change voice preset per request
- No re-deployment required
- Instant configuration updates
Monitoring usage metrics
Your dashboard tracks request metadata, including character counts and status, which is essential for billing accuracy and budget forecasting. Since each request is charged per character, monitoring usage helps you predict costs for large-scale text-to-speech deployments.
You can also review your key status and balance directly on the /get-api-key/ page to ensure uninterrupted service during high-traffic periods.
- Track character usage
- Monitor request status
- Check balance in real-time
Support and resources
Access comprehensive guides and API references directly through our documentation portal. The docs detail every endpoint, parameter, and code example needed for seamless integration into your accessibility tools.
Documentation links
Visit /docs/ for the complete API reference. It covers request structures, voice parameters, and error codes with ready-to-use curl examples.
- API Reference: /docs/
- Quick Start Guide
- Voice Parameter Details
Email support contact
For technical questions or billing issues, email us at [email protected]. We do not offer phone support or live chat; please include your API key and request ID for faster resolution.
- Response time: within 24 hours
- Include API key in subject
FAQ for voice selection
New to TTS? Check our FAQ for common questions on voice differences, language detection, and character limits. It helps you choose the right preset for your public information system.
- Voice vs. Language
- Character Limit Handling
- Trial Credit Usage
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Allow users to choose their preferred voice or disable audio entirely.
23 languages
Any voice speaks any of these languages.
