Frequently asked questions
Which TTS API is best?
A: There is no single best API; it depends on your specific needs. Auralux offers a simple, character-based TTS API with 8 voice presets and support for 23 languages. We focus on providing a straightforward integration for development teams without unnecessary complexity or proprietary constraints.
Is this API free?
No, but new accounts signing in with Google receive a $0.50 free trial valid for 7 days. Standard pricing is $10 per 1 million characters. Failed requests are free, and you can top up your balance using USDT on the TRON network with no minimum spend.
How many languages are supported?
We support 23 languages, including English, Spanish, German, French, Japanese, and Arabic. Any voice can speak any of these languages. If you do not specify a language, it is guessed from the script, defaulting to English for Latin text.
Can I clone voices?
No, we do not offer custom voice cloning. You can choose from 8 preset voices based on licensed recordings from LibriTTS-R and LibriVox. These voices include options like Aria, Nova, Felix, and Atlas, but you cannot upload your own voice or design custom ones.
Does it support SSML?
No, our API does not support SSML. You send plain text up to 4096 characters per request. If you need specific pronunciation or pauses, you can adjust the text directly or use the speed parameter to control the overall generation pace.
Is it real-time?
No, we do not offer streaming or WebSocket connections for real-time partial audio. You send a request and receive the full audio file once synthesis is complete. Generation takes roughly real-time duration, but requests are processed sequentially and may queue under load.
How does it compare to OpenAI?
Our API endpoint structure matches the OpenAI speech endpoint, so you can use the official OpenAI SDK by changing the base URL. However, we are not affiliated with OpenAI. We offer different voice presets, a different pricing model, and support for 23 languages.
What is the character limit?
Each request accepts up to 4096 characters of input text. Spaces and punctuation count toward this limit. For longer texts, you should split the text into parts, generate audio for each, and join the resulting files in your application logic.
Are there hidden fees?
No, there are no hidden fees. Pricing is $10 per 1 million characters. Failed requests are free. We do not charge for subscriptions, seats, or invoices. You only pay for successful character generation, and your balance does not expire.
Can I use it for commercial use?
Yes, you can use the generated audio for commercial purposes. The voices are based on licensed recordings (LibriTTS-R CC BY 4.0) and public-domain readings.
How do I get started?
Sign up at /get-api-key/ using Google or email. You receive an API key immediately. Use the key with the OpenAI SDK by setting the base URL to https://api.66mmr.info/v1. Send a POST request to /v1/audio/speech with your text and desired voice.
Is there a mobile SDK?
No, we do not provide a dedicated mobile SDK. Since our API follows the OpenAI request shape, you can use the official OpenAI SDKs for iOS, Android, or any platform by simply updating the base URL to our API endpoint.
What is the latency?
Latency varies based on load. Generation takes roughly real-time duration (25 characters per second). Short phrases take a few seconds. Requests are processed sequentially, so under heavy load, your request may wait in a queue before synthesis begins.
Do I need a credit card?
No, we do not accept credit cards or PayPal. You top up your balance using USDT on the TRON network (TRC-20). The minimum top-up is $5, and credits are added automatically within minutes. No subscriptions or invoices are required.
Key criteria
Selecting a text-to-speech API for voice assistants requires balancing cost, clarity, and coverage. Auralux simplifies this by offering straightforward per-character billing and consistent audio quality across its eight distinct voice presets.
Pricing model
We charge $0.01 per 1,000 characters, making it easy to predict costs for your bot’s dialogue. Failed requests are free, and there are no monthly subscriptions or hidden fees.
- Per-character billing ($0.01 per 1,000 chars)
- No monthly plans or minimum spend
- Failed requests are free
Voice quality
Our eight presets range from warm narrators to energetic promo voices, all based on licensed recordings. Each file includes an inaudible watermark to mark synthetic speech, ensuring transparency for your users.
- 8 distinct voice presets
- Licensed source recordings
- Inaudible synthetic watermark
Language support
We support 23 languages, allowing your assistant to converse globally. The API guesses the language from the script, but you can explicitly set it for non-English Latin text to ensure accuracy.
- 23 supported languages
- Auto-detection from script
- Explicit language override
Pricing comparison
Pricing for voice synthesis is straightforward: you pay only for what you process. There are no monthly subscriptions or minimum commitments, so your costs scale directly with usage. This model simplifies budgeting for large teams, as expenses are tied strictly to character count rather than unpredictable minute-based tiers.
Per-character billing
Each character, including spaces and punctuation, costs $0.00001. One million characters cost $10. Failed requests are free, ensuring you only pay for successful audio generation. This transparent unit price makes it easy to predict costs for high-volume chatbot deployments.
Free trial options
Signing in with Google provides $0.50 in trial credit, valid for 7 days. This allows teams to test the API with about 50,000 characters before committing. Email-based accounts start without this trial credit, requiring an immediate top-up to begin usage.
Payment methods
We accept USDT on the TRON network (TRC-20) for top-ups. Minimum deposits are $5, and balances are credited automatically within minutes. There are no credit cards, PayPal, or invoices with VAT, keeping the payment process simple and global.
Voice options
Auralux provides eight distinct voice presets engineered for voice assistants and chatbots. Each voice is selected for clarity and natural pacing, ensuring your application sounds professional without requiring complex configuration.
8 presets available
You can choose from eight specific voices, each optimized for different conversational contexts.
Gender variety
The selection includes four female and four male voices, allowing you to match your brand’s personality. This variety helps create a more engaging user experience by providing distinct auditory identities for different features or modes.
- Female: aria, nova, luna, sage
- Male: felix, milo, atlas, orion
Distinct personalities
Each voice carries a unique tone, from warm and storytelling to energetic and news-like. For example, 'nova' offers a bright, upbeat quality suitable for active engagement, while 'atlas' provides a neutral, authoritative tone ideal for information delivery.
Language support
Auralux supports 23 languages, enabling your voice assistant to communicate globally. You can mix languages within a single conversation, as any voice can speak any supported language. This flexibility ensures consistent voice identity across diverse user bases without needing to manage multiple voice models.
23 languages covered
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. This range covers most major markets, allowing you to deploy assistants that resonate with international audiences.
- Arabic, Chinese, Danish, Dutch, English
- Finnish, French, German, Greek, Hebrew, Hindi
- Italian, Japanese, Korean, Malay, Norwegian
- Polish, Portuguese, Russian, Spanish, Swahili
- Swedish, Turkish
Auto-detection
The API guesses the language from the script if you do not specify it. Latin script defaults to English. For non-English text using Latin characters, set the language field explicitly to ensure correct pronunciation and intonation.
Missing languages
Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, Bengali, and others are not yet supported. If your voice assistant targets these regions, you may need a fallback strategy or manual language tagging to handle unsupported inputs gracefully.
Technical features
Auralux uses a standard REST API endpoint that mirrors the OpenAI speech structure. You send a POST request with a JSON body containing the text, voice, and format. This design means your existing infrastructure works out of the box without custom adapters.
OpenAI SDK compatible
Since the request shape matches OpenAI’s speech endpoint, you can use official OpenAI SDKs by simply changing the base_url to https://api.66mmr.info/v1. Pass your tts-... key in the Authorization header, and the SDK handles the rest.
- Use any OpenAI SDK language binding
- Only change
base_urland auth header - Immediate integration for Python, JS, Node.js
REST API standard
The API follows standard HTTP conventions. You POST text to /v1/audio/speech and receive an audio file directly in the response. Standard JSON fields like model, input, voice, and response_format control the synthesis. This predictability makes debugging and logging straightforward for large development teams.
No custom SDK needed
We do not maintain a proprietary SDK. You rely on the standard HTTP client or the widely used OpenAI libraries. This avoids vendor lock-in and ensures you are not dependent on a single library’s update cycle. It also means your developers can use their preferred tools without learning a new SDK.
Performance
Auralux processes requests sequentially, typically generating audio at roughly real-time speed. Long texts are split into 4096-character chunks and stitched together seamlessly. Expect about 25 characters of text per second of generated audio for substantial content. Short phrases may take a few seconds to synthesize and return.
Under heavy load, requests queue behind others and process one by one. Your key allows 60 requests per minute once funded, but only 10 per minute before the first top-up. This linear processing ensures consistent quality, though you may experience slight delays if many developers are active simultaneously.
Real-time speed
Generation speed tracks closely with audio duration. You can expect roughly 25 characters of text per second of output for long-form content. Short phrases might take a few seconds to complete, as the engine initializes the audio pipeline for each request.
Queue handling
Requests are processed sequentially. If the API is busy, your request waits in a queue until a slot opens. This linear processing guarantees consistent audio quality but means higher concurrent usage may increase total generation time.
Rate limits
Your API key supports up to 60 requests per minute after topping up. During the free trial, this limit drops to 10 requests per minute. Exceeding these limits will result in standard rate-limit errors until the window resets.
- Standard limit: 60 requests/minute per key
- Trial limit: 10 requests/minute per key
- Processing is sequential, not parallel
Audio formats
Auralux supports five output formats, but MP3, WAV, and OPUS are the most common choices for voice assistant development. Selecting the right format balances file size, quality, and compatibility with your target hardware.
MP3 for web
MP3 is the default and most versatile choice for web-based voice assistants. It offers a good balance between audio quality and file size, ensuring fast loading times for users with varying connection speeds.
WAV for quality
Choose WAV when you need uncompressed, high-fidelity audio. This format is ideal for local playback on devices that prioritize audio clarity over storage constraints, preserving the full range of the TTS engine.
OPUS for size
OPUS provides superior compression for bandwidth-constrained environments. It delivers excellent quality at lower bitrates, making it perfect for mobile apps or voice assistants operating on limited data plans.

Integration ease
Integration is straightforward because the API follows a standard REST pattern compatible with existing tools. You can send requests using any HTTP client or the official OpenAI SDK by simply pointing it to our base URL. This means your team doesn’t need to learn a new library or manage a custom SDK to get started quickly.
The JSON body structure is simple: specify the model, input text, voice, and format. You handle the rest with standard code. Even error handling is predictable, returning clear status codes so you can retry failed requests; failed requests are free.
Simple curl example
You can test the endpoint immediately with a single command. This approach works for any language or framework that supports HTTP requests.
curl -X POST 'https://api.66mmr.info/v1/audio/speech' \ -H 'Authorization: Bearer tts-...' \ -H 'Content-Type: application/json' \ -d '{"model": "tts-1", "input": "Hello world", "voice": "nova", "response_format": "mp3"}'
JSON body format
The request body accepts standard fields like model, input, voice, and response_format. You can also adjust speed between 0.25 and 4.0 or specify a language ISO code if needed. This structure mirrors the OpenAI speech endpoint, making it easy to swap or integrate.
Just ensure your text stays under 4096 characters per request. If you have longer dialogues, split the text and join the resulting audio files in your application logic.
- Model: tts-1 or tts-1-hd
- Voice: e.g., nova, aria, atlas
- Format: mp3, wav, opus, flac, pcm
- Speed: 0.25 to 4.0
Error handling
If synthesis fails, the API returns an error code and refunds the characters instantly. This ensures you never pay for broken audio. Your code can catch these errors and retry or alert the user without losing budget.
For rate limits, you’ll receive a 429 status if you exceed 60 requests per minute. Implementing simple exponential backoff in your chatbot logic handles this gracefully during peak usage.
Data privacy
Auralux treats your data with straightforward transparency. When you send text for synthesis, it is used exclusively to generate the audio file you receive. We do not use your input text to train our models or improve the engine, ensuring your content remains private and yours.
No training on input
Your voice data stays yours. The text you provide is processed solely for the immediate generation of the audio response. It is not ingested into our training datasets, so your unique scripts or conversations won’t influence future model outputs without your explicit consent.
No audio storage
Generated audio files are returned in the response stream and are not permanently stored on our servers. Once the audio is delivered to your application, we do not keep a copy of the resulting speech file, minimizing your exposure to data retention risks.
Metadata only
We retain only the essential metadata required for accurate billing and service monitoring. This includes the timestamp, character count, and request status. This minimal data set allows us to manage your account and rate limits efficiently without hoarding your conversation history.
Limitations
While Auralux delivers high-quality synthetic speech, it is designed as a straightforward API rather than a full media player. Understanding these boundaries helps you architect your voice applications correctly without unexpected friction during integration.
No streaming
The API returns a complete audio file after synthesis finishes. You do not receive partial chunks or real-time audio streams, which means your frontend must handle buffering or wait for the full response before playing.
No SSML
The service does not support Speech Synthesis Markup Language (SSML). You cannot control pauses, emphasis, or pronunciation rules via XML tags. Voice control is achieved simply by selecting one of the eight preset voices.
No voice cloning
Auralux uses eight licensed presets derived from LibriTTS-R and LibriVox recordings. There is no option to upload a custom voice sample or clone a specific person’s voice, keeping the catalog stable and predictable.
Who is it for?
Auralux is built for engineering teams who need reliable voice synthesis without managing complex infrastructure. Our character-based pricing and OpenAI-compatible SDK let you integrate speech into applications quickly, handling the audio generation so you can focus on conversation design and user experience.
Chatbot developers
Integrate natural-sounding responses into conversational flows with minimal code. The standard REST API and familiar JSON body format allow you to drop speech generation into existing chatbot logic without learning a proprietary SDK.
- Compatible with standard OpenAI SDKs
- Simple JSON request structure
- Supports 23 languages for global reach
Voice app teams
Build voice assistants that respond with audio generated in roughly real-time. With realistic speeds and eight distinct voice presets, you can create engaging audio experiences that feel responsive. The system handles queue management, ensuring consistent delivery even under load.
- Eight unique voice personalities
- Realistic generation speeds
- Support for MP3, WAV, and OPUS formats
Accessibility projects
Provide clear, synthetic speech for visually impaired users or content consumption. Auralux ensures input text is not used for training, preserving user privacy while delivering high-quality audio output directly in the response.
- No training on input data
- No audio storage after generation
- Multiple audio formats for compatibility
Final verdict
Auralux offers a straightforward API designed for development teams who want to integrate voice capabilities without managing complex audio infrastructure. The character-based pricing model ensures that costs scale linearly with usage, making it easy to predict expenses for both small prototypes and large-scale voice assistants.
Best for simplicity
The API mirrors the OpenAI speech endpoint, allowing you to use familiar SDKs with a simple base_url change. This reduces integration time significantly, letting your team focus on conversation design rather than debugging custom protocols or managing proprietary libraries.
- Standard JSON request body
- No custom SDK required
- Direct audio file response
Cost-effective scaling
With billing at $10 per million characters and no monthly fees, you only pay for what you use. Failed requests are free, and the USDT top-up process allows for flexible, instant credit loading without waiting for invoice cycles or subscription renewals.
- $0.01 per 1,000 characters
- No minimum spend
- Instant USTRC-20 top-ups
Developer friendly
Support is handled via email, providing a clear channel for technical questions without the noise of live chat. The API returns standard HTTP status codes, and the consistent request shape makes error handling predictable for engineering teams building robust voice-enabled applications.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This model simplifies budgeting for large teams, as expenses are tied strictly to character count rather than unpredictable minute-based tiers.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- LibriVox public-domain audiobooks
- Tether (USDT)
- ISO 639 language codes
- LibriTTS-R corpus (OpenSLR 141)
- openai-node on GitHub
- IETF RFC 7845: Ogg Opus
- MDN: Web Speech API
- Creative Commons Attribution 4.0
- LibriTTS-R paper (arXiv 2305.18802)
- FFmpeg documentation
- W3C WAI: making audio and video accessible
