Frequently asked questions
What is text-to-speech?
Text-to-speech (TTS) is a cloud API that converts written text into audio files. You send text via a POST request to our endpoint, and receive an audio file in formats like mp3, wav, opus, flac, or pcm. It is a simple integration for generating synthetic speech without manual recording.
Why use TTS in restaurants?
TTS provides reliable audio for menus, notifications, and announcements without hiring voice actors. It allows instant updates to menu items or prices. The character-based pricing keeps costs predictable, and the lightweight API integrates easily with existing restaurant tech stacks for digital signage or drive-thru systems.
How much does it cost?
Pricing is $10 per 1 million characters, or $0.01 per 1,000 characters. Spaces and punctuation count as characters. Failed requests are free. There are no monthly plans, seats, or minimum spends. Top-ups via USDT on TRON are credited automatically within minutes.
Which voice is best for menus?
For menus, aria (warm, clear narrator) or sage (calm, instructional) are ideal. Felix works well for energetic promo announcements. Each voice is a preset based on licensed recordings. You can switch voices instantly by updating the voice parameter in your API request without re-recording audio.
Can I use multiple languages?
Yes, the API supports 23 languages including English, Spanish, French, German, Japanese, and Chinese. Any voice can speak any supported language. Set the language field to an ISO code for non-English Latin text to ensure accurate pronunciation, as the system guesses the language otherwise.
How do I play the audio?
The API returns an audio file in the response. You can save this file to your server or play it directly in a web player. Supported formats include mp3, wav, opus, flac, and pcm. Integrate the audio URL or file into your restaurant’s digital display or app.
What if the audio fails?
If synthesis fails, the charge is returned in full. You can retry the request. For longer texts, split them into parts of up to 4096 characters and join the resulting audio files. Failed requests do not incur costs, ensuring you only pay for successful generation.
Do I need a developer?
A basic developer is helpful for integrating the REST API, but it is straightforward. You can use the official OpenAI SDKs by changing the base_url. The endpoint accepts standard JSON bodies. Simple curl examples are provided on the /get-api-key/ page for quick testing.
Can I update menus in real-time?
Yes, you can generate new audio instantly when menu items change. Since text is converted to audio on request, updates are immediate. You can trigger a new API call whenever a dish is added or removed, ensuring your digital menus always reflect current availability.
Is the audio copyrighted?
Voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. Generated files carry an inaudible watermark marking them as synthetic speech. The audio is returned in the response and not stored by us, giving you full control over the generated files.
How do I handle special characters?
Send the text as a standard JSON string in the input field. Special characters are treated as part of the text input. Spaces and punctuation count toward the character limit. The API processes the text directly, so no special escaping beyond standard JSON formatting is needed.
What is the latency?
Typical speed is roughly real time, about 25 characters per second. Short phrases take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue. Plan for slight delays during high traffic periods on the API.
Can I use this for drive-thrus?
Yes, the API is suitable for drive-thru announcements. Generate audio for order confirmations or menu items on demand. The lightweight nature of the API allows integration with existing drive-thru hardware or software. Ensure your system handles the audio playback format supported by your speakers.
How do I track usage?
The dashboard on /get-api-key/ shows your balance, characters left, and usage history. Request metadata, including time, characters, and status, is kept for billing. You can monitor your consumption directly from the account page to manage your USDT top-ups effectively.
Is there a free trial?
Yes, signing in with Google gives $0.50 of credit, valid for 7 days, for about 50,000 characters. Email/password accounts start without trial credit. Top-ups via USTRC20 are required for continued use after the trial or initial balance is exhausted.
Do I need to pay for retries?
No, failed requests are free. If synthesis fails, the charge is returned in full. You can retry the request without additional cost. This ensures you only pay for successfully generated audio files, protecting your budget during integration testing or high-load scenarios.
Can I change voices later?
Yes, you can change the voice for any new request by updating the voice parameter. There is no need to regenerate existing audio files. You can switch between the 8 presets, such as from aria to felix, instantly for new menu items or announcements.
What is the maximum text length?
The maximum input text length is 4096 characters per request. For longer menus, split the text into parts of this limit and join the resulting audio files. This allows you to generate audio for extensive menus or long announcement scripts efficiently.
Does the API work offline?
No, this is a cloud API. You must send text to the endpoint https://api.playak56.com/v1/audio/speech to receive audio. It requires an internet connection. There is no offline mode or local SDK. Audio is generated on our servers and returned in the response.
Is data stored after generation?
Input text is used only to produce the audio and is not used for training. Generated audio is returned in the response and not stored by us. Request metadata (time, characters, status) is kept for billing purposes. Your data remains under your control.
Understanding TTS in Hospitality
Text-to-speech (TTS) converts written text into natural-sounding audio files via a cloud API. For restaurant tech startups, this enables automated audio menus, order notifications, and ambient announcements without hiring voice talent or managing complex studio equipment.
The API processes requests as simple HTTP calls, returning audio in formats like MP3 or WAV. This lightweight integration allows dining halls to broadcast clear, consistent audio for daily specials or wayfinding, enhancing the customer experience with minimal technical overhead.
Definition of TTS
TTS synthesizes speech from text input, supporting 23 languages including English, Spanish, French, and German. The engine generates audio files directly from your application, ensuring reliable playback across web interfaces and digital signage systems.
Benefits for Restaurants
Automated audio updates reduce labor costs for menu changes. Consistent voice quality maintains brand identity across multiple locations. Real-time generation allows for instant updates to daily specials or dietary alerts without pre-recording.
Character-based pricing ($0.01 per 1,000 characters) ensures predictable costs. Failed requests are free, protecting your budget during peak hours. The system handles up to 4,096 characters per request, simplifying content management.
- Instant menu updates via API
- Consistent voice across locations
- Predictable character-based pricing
- No monthly subscription fees
Common Use Cases
Digital menu boards can read aloud daily specials or allergen warnings. Order confirmation systems can verbally confirm items to customers. Waiting area announcements can notify guests of table availability or reservation times.
Kitchen display systems can use audio cues for order readiness. Multi-language support allows menus to be read in tourist-preferred languages, improving accessibility for international visitors without requiring separate audio files for each language.
- Digital menu boards
- Order confirmation alerts
- Wait time announcements
- Multi-language accessibility
Setting Up the API
To begin using SavorVoice for your restaurant tech stack, start by visiting the /get-api-key/ page. You can sign up instantly using your Google account or create a standard email and password login. This account is specific to playak56.com and separates your usage data from other potential integrations.
Once registered, you will receive a unique API key. This single key authenticates all your requests. You can generate a new key at any time, which immediately invalidates the old one, ensuring secure access across your dining halls or mobile apps without downtime.
Create an Account
Navigate to /get-api-key/ to register. Google sign-in provides an immediate $0.50 trial credit valid for 7 days, allowing you to test synthesis without upfront payment. Email accounts start with zero balance, requiring a top-up before use.
Generate API Key
After signing up, your dashboard displays your active API key. This key starts with 'tts-'. You can issue a new key at any moment; the previous key stops working immediately, simplifying security management for staff turnover or app updates.
Test Integration
Use the provided curl example on your dashboard to verify connectivity. Send a POST request to https://api.playak56.com/v1/audio/speech. Ensure your base URL is set correctly if using the OpenAI SDK. A successful response returns an audio file, confirming your integration is live.
- Use POST https://api.playak56.com/v1/audio/speech
- Include Authorization: Bearer tts-<your-key>
- Check response for audio data or errors
Choosing the Right Voice
Selecting the correct voice preset ensures your audio aligns with the specific context of your restaurant’s operations. Each voice offers a distinct tonal quality, from warm narration to neutral reporting, allowing you to tailor the listener’s experience for menus, customer interactions, or operational updates.
Matching voice personality to function reduces cognitive load for diners and staff. Aria provides a clear, warm narrator style ideal for reading dishes. Milo offers a friendly, conversational tone suitable for queue notifications. Atlas delivers a neutral, news-style delivery perfect for clear, unobtrusive announcements.
Aria for Menus
Use Aria (warm, clear narrator) for digital menu displays or table talkers. Her tone is inviting and easy to understand, making complex dish descriptions accessible without sounding robotic. This voice works best for static descriptions where clarity is paramount.
Aria’s calm delivery ensures that allergen warnings or dietary notes are heard clearly. Since menus often contain specific ingredients, her clear enunciation helps prevent misinterpretation. Use her for short, descriptive texts up to 4096 characters per request.
- Ideal for: digital menu boards, table talker audio, dish descriptions
Milo for Interactions
Deploy Milo (friendly, conversational) for customer-facing interactions such as queue management or table readiness alerts. His tone is approachable, reducing the stiffness often associated with automated systems. This voice feels more like a staff member than a machine.
Milo is effective for brief, interactive prompts like "Your table is ready" or "Please proceed to the counter." The conversational style helps maintain a welcoming atmosphere even during high-traffic periods. Ensure text is concise to keep the interaction natural.
- Best for: queue notifications, table readiness alerts, greeting messages
Atlas for Updates
Choose Atlas (neutral, news) for operational updates or staff announcements. His neutral tone conveys information objectively without emotional bias, which is ideal for time-sensitive or factual messages. This voice is unobtrusive and professional.
Use Atlas for daily specials, kitchen status updates, or emergency announcements. His clear, news-style delivery ensures the message is prioritized and understood quickly by staff or diners. This voice works well for repetitive, informational content where tone consistency matters.
- Recommended for: staff announcements, daily specials, emergency alerts
Technical Implementation
To generate speech, send a POST request to https://api.playak56.com/v1/audio/speech with your JSON payload. The API accepts standard OpenAI SDK parameters, allowing you to switch the base URL and use your Bearer token. You specify the model, input text, voice, and format. For texts exceeding 4,096 characters, you must split the input yourself, generate audio for each chunk, and join the files manually.
HTTP Requests
Use the official OpenAI SDK or plain HTTP. Set the base URL to https://api.playak56.com/v1 and pass your key in the Authorization header. Each request must include the text input, voice preset, and desired audio format. The API processes requests sequentially; under load, you may experience queue delays.
- Endpoint: POST /v1/audio/speech
- Auth: Bearer token in header
- Format: JSON body
- Limit: 4096 chars per request
Audio Playback
The API returns audio files directly in the response. Since it does not stream audio, you must handle playback locally or via your web server. For short phrases, generation takes a few seconds. For longer menus, split the text, generate each part, and join the audio files to ensure continuous playback.
- No streaming support
- Manual file joining required
- Real-time generation speed
Error Handling
If synthesis fails, the charge is returned in full. Always check the HTTP status code. Rate limits are strict: 60 requests per minute per key. If you exceed this, wait or implement exponential backoff. Failed requests do not consume your character balance.
- Failed requests are free
- 60 requests/minute limit
- Check status codes
Managing Character Limits
Text-to-speech requests accept up to 4,096 characters per call. For longer menu descriptions or announcement scripts, split the input into smaller chunks and join the resulting audio files to maintain coherence.
This approach ensures reliable synthesis without hitting request limits, keeping your audio delivery smooth for customers at digital kiosks or drive-thru displays.
Each generated segment is stitched seamlessly, preserving the natural flow of your restaurant's voice. This method works efficiently for both short phrases and extended daily specials.
The API handles the splitting logic, allowing you to focus on content rather than technical constraints. This keeps your integration simple and robust.
By monitoring character usage, you can optimize your scripts to stay within the 4,096-character limit per request, minimizing overhead and ensuring faster response times.
This strategy helps you maintain high-quality audio output while managing costs effectively, as each character counts toward your billed usage.
Text Splitting
Requests are limited to 4,096 characters. Split longer texts into parts, generate each audio file, and join them. This ensures reliable synthesis without hitting request limits.
The API handles the splitting logic, allowing you to focus on content rather than technical constraints. This keeps your integration simple and robust.
By monitoring character usage, you can optimize your scripts to stay within the 4,096-character limit per request, minimizing overhead and ensuring faster response times.
- Split text into chunks of <= 4096 characters.
- Generate audio for each chunk separately.
- Join the resulting audio files seamlessly.
Cost Optimization
Pricing is $0.01 per 1,000 characters, including spaces and punctuation. Efficient text management reduces your monthly spend significantly.
Use concise language in your menu descriptions and announcements to maximize the value of each character. Shorter scripts mean lower costs.
Monitor your usage via the dashboard to track character consumption. This helps you identify areas for optimization and avoid unnecessary expenses.
- Pricing: $0.01 per 1,000 characters.
- Include spaces and punctuation in character count.
- Monitor usage via the dashboard.
Quality Control
Ensure each split segment maintains the intended tone and pacing. Test the joined audio for natural flow and clarity.
Use the 'aria' voice for menus and 'milo' for interactions to maintain a consistent brand voice across all audio outputs.
Regularly review the generated audio to ensure it meets your quality standards. Adjust the script or voice settings as needed.
- Maintain tone and pacing across segments.
- Use consistent voice settings for brand identity.
- Regularly review generated audio for quality.
Language Support
Our API supports 23 languages, enabling restaurants to serve diverse customer bases without complex integrations. Each voice can speak any supported language, ensuring consistent audio quality across different regions. This flexibility is crucial for establishments with multilingual patrons or international tourist traffic.
Text is processed using ISO language codes or automatically detected from the script. Setting the language explicitly is recommended for non-English Latin text to avoid misinterpretation. Unsupported languages like Vietnamese or Thai will not be processed correctly.
Multi-Language Menu
Support for 23 languages allows you to display menus in Arabic, Chinese, French, German, Hindi, Japanese, Korean, Russian, Spanish, and more. Each voice maintains its distinct tone regardless of the language spoken, ensuring a cohesive brand experience.
Customers can listen to daily specials or allergen information in their preferred language. This reduces wait times and improves accessibility for non-native speakers without hiring additional staff.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Tourist-Friendly Audio
Tourist-heavy locations benefit from clear audio in multiple languages. Use the API to generate announcements for table reservations, order readiness, or closing times in languages common among your visitors.
The automatic language detection works well for standard texts. However, for mixed-language menus, explicitly setting the language parameter ensures accurate pronunciation and avoids defaulting to English for Latin-script texts.
Local Dialects
While we support 23 languages, we do not offer specific dialect variations or custom voice cloning. Each language is handled with standard pronunciation models. For example, 'English' covers standard US/UK pronunciations, but regional dialects are not distinct.
- No custom dialect support
- Standard pronunciation per language
Audio Formats for Devices
Choosing the right audio format balances quality, compatibility, and bandwidth for restaurant hardware. The API supports MP3, WAV, OPUS, FLAC, and PCM, allowing you to match the output to your specific playback device.
MP3 remains the standard for web-based menu players and mobile apps due to its universal support. WAV is ideal for legacy POS systems that require uncompressed PCM audio. OPUS offers significant bandwidth savings for high-volume announcements without sacrificing clarity.
MP3 for Web
Use MP3 for digital menus displayed on tablets or web interfaces. It provides a good balance of quality and file size, ensuring fast loading times for customers browsing options.
Most modern browsers and media players handle MP3 natively, reducing the need for additional decoding libraries in your frontend stack.
- Universal compatibility with web players
- Small file size for quick loading
- Standard format for digital signage
WAV for Systems
WAV delivers uncompressed audio, which is preferred by many legacy Point-of-Sale (POS) terminals and in-kitchen display systems. It ensures no quality loss during playback on hardware that may not support compressed formats.
This format is best when you need guaranteed fidelity for critical alerts or announcements where every word must be heard clearly without compression artifacts.
- Uncompressed audio quality
- Ideal for legacy POS hardware
- No decoding overhead
OPUS for Efficiency
OPUS provides up to 50% bandwidth savings compared to MP3 while maintaining clear voice quality. It is ideal for high-volume announcements sent over cellular or constrained networks to kitchen displays or outdoor speakers.
Use OPUS when you need to minimize data usage for frequent, short audio updates across multiple devices without compromising intelligibility.
- Up to 50% bandwidth savings vs MP3
- Ideal for high-volume announcements
- Maintains clear voice quality

Monitoring Usage
Tracking your API usage is essential for managing operational costs in a restaurant environment. The dashboard provides real-time visibility into your account status, ensuring you never exceed your budget unexpectedly.
Check Balance
Access your current balance directly from the /get-api-key/ page. Each account holds a single active key, and any unused credit remains available until consumed. Failed requests are refunded immediately, so your balance reflects only successful synthesis.
- Real-time balance updates
- Automatic refund for failed requests
- Single active key per account
Track Requests
View detailed logs of each request, including timestamp, character count, and status. This data helps identify peak usage times in your dining halls or kitchen displays.
- Request timestamp logging
- Character count per request
- Success/failure status indicators
Optimize Costs
Since pricing is $0.01 per 1,000 characters, precise text management matters. Avoid redundant phrasing in menu descriptions or announcements to maximize the value of your $10 per million character credits.
- Cost per 1,000 characters
- Minimal overhead for short phrases
- No minimum spend requirements
Common Issues
Text-to-speech in hospitality relies on sequential processing, meaning requests are handled one after another. During peak hours, your audio generation may queue behind other tasks, causing slight delays in playback. This is normal behavior for our infrastructure, not a system failure.
Understanding these technical nuances helps you design robust user experiences that account for minor latency without frustrating your diners or staff.
Audio Delays
Requests are processed sequentially, so under load a request may wait in a queue. Longer texts take longer to synthesize, roughly real-time at 25 characters per second. Short phrases may still take a few seconds due to processing overhead.
If your system expects instant playback, buffer the audio or provide visual feedback to users while the synthesis completes in the background.
- Expect slight queuing during high traffic
- Short phrases take a few seconds
- Plan for sequential processing
Character Count Errors
Each request accepts a maximum of 4096 characters. If your menu description or announcement exceeds this limit, the API will reject the request. You must split the text into chunks and join the resulting audio files manually.
Spaces and punctuation count toward the character total, so plan your text length accordingly to avoid unexpected billing or truncation issues.
- Max 4096 characters per request
- Include spaces and punctuation in count
- Split long texts before sending
Voice Mismatch
Voices are language-agnostic but may sound unnatural if the language parameter is omitted for non-English Latin text. The system guesses the language from the script, defaulting to English for Latin characters.
Always explicitly set the language field for non-English content to ensure correct pronunciation and intonation from voices like Aria or Milo.
- Set language explicitly for non-English
- Default guess may be incorrect
- Use ISO codes for clarity
Best Practices
Optimizing audio output for restaurant environments requires careful attention to text structure and delivery consistency. Since the API processes requests sequentially, short, distinct phrases reduce queue wait times and ensure immediate playback for time-sensitive announcements.
Maintaining a uniform voice across all digital touchpoints, from digital menus to table notifications, builds brand recognition without distracting diners. Clear, concise inputs also help the model avoid awkward pauses or mispronunciations.
Short Sentences
Break long menu descriptions into separate API calls. Each request supports up to 4,096 characters, but shorter inputs (under 100 characters) generate faster, reducing latency for interactive systems like table talkers.
- Limit each input to a single sentence or phrase.
- Avoid complex punctuation that might confuse the parser.
Clear Pronunciation
Use standard spelling for dish names and ingredients. The engine guesses language from script, so explicitly setting the 'language' parameter prevents mispronunciation of non-English terms, ensuring accurate delivery for international guests.
- Spell phonetic names clearly (e.g., 'sushi' not 'su-shi').
- Set the 'language' field explicitly for mixed-script text.
Consistent Tone
Stick to one voice preset per location. Using 'Milo' for interactions or 'Aria' for menus creates a predictable auditory experience. Avoid switching voices mid-journey to prevent jarring transitions for the customer.
- Assign specific voices to specific use cases.
- Maintain the same speed setting (e.g., 1.0) across all audio.
Scaling for Growth
Scaling your restaurant’s audio infrastructure requires handling multiple locations and fluctuating traffic without compromising delivery speed or incurring unexpected costs. Our character-based pricing model ensures that expanding to new venues or increasing notification frequency remains predictable and affordable, regardless of volume.
Multiple Locations
Each location can operate with its own API key, allowing you to isolate usage and billing per branch. This structure simplifies cost attribution and ensures that one location’s high traffic does not impact the rate limits of another.
- Independent keys per location for clear billing
- Consistent voice quality across all venues
- No additional infrastructure costs for multi-site deployment
High Traffic
With a base limit of 60 requests per minute, you can handle peak dining hours efficiently. For higher volumes, top-up your balance via USDT to maintain continuous operation, as requests are processed sequentially and may queue under heavy load.
- Base 60 requests/minute per key
- Sequential processing with queue handling
- Instant balance updates via TRC-20 transfers
Automated Updates
Integrate the API into your POS or menu management system to push audio updates automatically when prices or items change. Since text is converted to audio on demand, you avoid storing large audio files and ensure customers always hear the latest menu information.
- Real-time menu updates via API
- Reduced storage needs with on-demand generation
- Consistent customer experience with fresh content
Support and Contact
We do not offer phone lines, live chat, or SLAs. For restaurant tech teams needing reliable audio integration, support is handled exclusively via email. This keeps our pricing simple and transparent, without the overhead of enterprise-grade guarantees or 24/7 staffing.
Our documentation is the primary resource for technical implementation. It includes curl examples and endpoint details for the /v1/audio/speech endpoint. We recommend reviewing the rate limits and character counting rules before deploying to production environments.
Email Support
Send questions about billing, API keys, or integration issues to our support team. We respond to emails during standard business hours. There is no ticket queue system; your email is processed in the order received.
- No phone support available
- No live chat widget
- Email only for all inquiries
Documentation
The docs cover all endpoints, including voice presets and language codes. You will find ready-to-use curl commands for generating speech. The guide explains how to handle the 60 requests per minute limit and how to split long menu texts.
- OpenAPI spec available
- Curl examples included
- Voice preset descriptions
Community
We do not maintain a dedicated community forum or Discord server. Developers should refer to the code examples in the documentation. For specific technical questions, email support is the only channel.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Your input text is used exclusively to generate the requested audio file and is not retained for model training.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- LibriTTS-R paper (arXiv 2305.18802)
- OpenAI API reference: create speech
- IETF RFC 7845: Ogg Opus
- Node.js documentation
- LibriVox public-domain audiobooks
- Tether (USDT)
- W3C WAI: making audio and video accessible
- openai-python on GitHub
- ISO 639 language codes
- Creative Commons Attribution 4.0
- TRON developer hub: TRC-20
