Frequently asked questions
Is the API suitable for reading long texts like Esperanza Rising?
Yes. The API handles long texts by splitting them into parts of up to 4096 characters and joining the audio files. This ensures complete coverage of lengthy books like Esperanza Rising.
Which voice is best for dyslexic readers?
Q: Which voice is suitable for dyslexic readers? A: Aria (warm, clear narrator) and Luna (soft, storytelling) are often preferred for readability. However, all eight presets are available for testing. Since individual preferences vary, we recommend evaluating which voice provides the clearest audio experience for your specific user base rather than assuming a single optimal choice.
How does the API handle text longer than 4096 characters?
The API accepts input text of up to 4096 characters per request. You must split longer texts and join the resulting audio files.
What formats are supported for audio output?
The API supports mp3, wav, opus, flac, and pcm formats. You can specify the desired response format in the request body. This allows flexibility for different playback environments and accessibility tools used by your readers.
Does the API support speed adjustment for slower readers?
Yes. You can adjust the playback speed between 0.25 and 4.0. This allows you to slow down the audio for readers who need more time to process text, enhancing accessibility for users with dyslexia or processing delays.
Is the input text stored after synthesis?
No. 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. Only request metadata is kept for billing purposes.
Can I use the OpenAI SDK with this API?
Yes. The request shape matches the OpenAI speech endpoint. You can use the official OpenAI SDKs by changing the base_url to https://api.6zzz.info/v1 and using your site's API key. No custom SDK is required.
What happens if a request fails?
Failed requests are free. If synthesis fails, the charge is returned in full. This ensures you are not billed for unsuccessful audio generation attempts, protecting your account balance during development or high-load scenarios.
Are there any monthly subscriptions?
No. There are no monthly plans, no seats, and no minimum spend. You pay only for characters used. Balances do not expire, and you can top up via USDT on the TRON network when needed.
How many languages are supported?
Twenty-three languages are supported, including English, Spanish, French, German, and others. Any voice can speak any of these languages. Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, and Bengali are not supported.
Is there a free trial available?
Yes. Signing in with Google provides $0.50 of credit, valid for 7 days, covering about 50,000 characters. Accounts created with email and password start without trial credit. This allows testing before committing to top-ups.
How do I authenticate my requests?
Use the Authorization header with the value Bearer followed by your API key. You receive one active key per account. You can issue a new key at any time, which immediately invalidates the old one.
What is the rate limit for requests?
The limit is 60 requests per minute per key. Before the first top-up, the limit is 10 requests per minute. Requests are processed sequentially, so under load, a request may wait in a queue.
Do I need to specify the language for every request?
No, but it is recommended for non-English Latin-script text. If not set, the language is guessed from the script, defaulting to English for Latin text. For other languages, explicitly setting the ISO code ensures correct synthesis.
First request
curl
curl https://api.6zzz.info/v1/audio/speech \
-H "Authorization: Bearer $TTS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"tts-1","voice":"aria","input":"Hello from the API.","response_format":"mp3"}' \
--output hello.mp3API Capabilities for Reading Text
AudioLex Access provides a cloud API for text-to-speech conversion, enabling developers to integrate audio synthesis directly into accessibility applications. The service accepts text input and returns generated audio files via a standard HTTP endpoint, allowing seamless integration with existing reading infrastructure without requiring complex custom middleware.
JSON-based POST endpoint
The API accepts text input via a JSON body sent to the POST /v1/audio/speech endpoint. Developers send the text, voice selection, and format preferences in a single request. The system processes the input and returns the resulting audio file, maintaining compatibility with standard HTTP workflows used by most modern development stacks.
Server-side synthesis
Text is converted to audio on the server side. The process delivers complete audio files efficiently. Requests are processed sequentially; under load, a request may wait in a queue. Typical generation speed is roughly real-time, producing approximately 25 characters of text per second of audio for long texts, while short phrases may take a few seconds.
Support for standard formats
The API supports multiple audio output formats including MP3, WAV, OPUS, FLAC, and PCM. Developers can specify the desired format in the request body. This flexibility allows integration with various player implementations and ensures compatibility with diverse accessibility tools and reading applications used by end-users.
Voice Selection for Narrative Clarity
Selecting the appropriate voice model ensures that text-to-speech output meets specific accessibility standards for different reading contexts. Our API provides distinct preset voices optimized for narrative clarity, instructional precision, and conversational engagement.
Each voice provides a distinct tone, allowing developers to tailor the listening experience to the content type.
Warm narrator voice: Aria
Aria delivers a warm, clear narration style ideal for literary works like 'Esperanza Rising.' Her tone supports immersive storytelling, helping users maintain focus on complex narrative structures.
As a female preset based on licensed LibriTTS-R recordings, Aria provides a consistent auditory anchor for chapter-by-chapter reading sessions.
- Optimized for narrative flow
- Clear articulation of punctuation
- Suitable for long-form text
Clear instructional tone: Sage
Sage offers a calm, instructional voice perfect for educational materials or guide-based content. Her neutral delivery minimizes emotional distraction, aiding comprehension for users with reading difficulties.
Use Sage when the primary goal is information retention rather than entertainment.
- Neutral pacing
- High clarity
- Low emotional variance
Conversational style: Milo
Milo provides a friendly, conversational tone that makes technical or dense text feel more approachable. His style is effective for interactive applications or user guides where a human touch is desired.
Like all voices, Milo supports all 23 languages and can be adjusted for speed between 0.25x and 4.0x to match user preference.
- Approachable tone
- Supports speed adjustment
- Ideal for UI feedback
Language Support for Diverse Readers
AudioLex Access supports 23 languages, including English, Spanish, French, German, Japanese, and Arabic. This range allows developers to build applications that serve diverse reader communities without requiring multiple distinct API implementations.
23 supported languages
The API handles 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 preset can pronounce any of these languages correctly, ensuring consistent audio quality across the supported linguistic set.
- 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 system guesses the language from the input script. If the language field is omitted, Latin script text defaults to English. For non-English Latin-script text, you should set the language parameter explicitly to avoid incorrect pronunciation.
Explicit language code setting
Set the language field to an ISO code to ensure accurate synthesis. This prevents the default English assumption for Latin-script text that is not English. Supported languages do not include Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, or Bengali.
Pricing Structure for Developers
AudioLex Access charges $0.01 per 1,000 characters, including spaces and punctuation. Failed requests receive a full credit refund, ensuring you pay only for successful synthesis. This transparent model aligns with the needs of indie developers building accessibility tools.
There are no monthly subscription fees or minimum spend requirements. You retain control over your budget without recurring commitments. The balance never expires, allowing you to top up only when necessary for ongoing development cycles.
$0.01 per 1,000 characters
The rate applies to all successful requests. Input text up to 4096 characters per request is billed proportionally. This predictable cost structure simplifies budgeting for long-form content like chapter-by-chapter audio generation.
No monthly subscription fees
Unlike traditional SaaS models, there are no seats or recurring base costs. You are not locked into a plan. This flexibility supports projects with variable usage patterns, common in accessibility app development.
Pay-as-you-go model
Top-ups are processed via USDT on the TRON network (TRC-20) with a $5 minimum. Credits appear automatically within minutes. This streamlined payment method avoids invoices and VAT complications, accelerating integration for global developers.
- USDT (TRC-20) only
- Minimum top-up: $5
- No PayPal or credit cards
- Balance does not expire
Technical Specifications
The API accepts text payloads up to 4096 characters per request. For longer works like "Esperanza Rising," clients must split the narrative into segments and concatenate the resulting audio files. Requests are processed sequentially, so under load, synthesis may queue behind prior jobs.
4096 character input limit
Each POST request supports a maximum of 4096 characters, including spaces and punctuation. Input exceeding this limit requires client-side segmentation. The API does not handle chunking automatically; the developer must manage the splitting logic and subsequent audio joining.
Speed adjustment 0.25x to 4.0x
The speed parameter allows fine-tuning of playback rate between 0.25 and 4.0. Standard generation approximates real-time speech, processing roughly 25 characters per second. Slower speeds may assist readers with processing delays, while faster rates suit quick review.
MP3, WAV, OPUS, FLAC outputs
Audio is returned in the requested format via response_format. MP3 and WAV provide broad compatibility, while OPUS and FLAC offer efficient or lossless storage. PCM is available for raw data. All generated files include an inaudible watermark identifying them as synthetic speech.
- MP3: Standard compressed format
- WAV: Uncompressed audio
- OPUS: Efficient compression
- FLAC: Lossless quality

Data Privacy and Usage
We process your input text exclusively to generate the requested audio file. The content is not retained for model training or improved by usage, ensuring your proprietary narratives remain private.
Once the audio is generated and returned in the response, the file is not stored on our servers. This minimizes data exposure and aligns with strict privacy requirements for educational materials.
Text not used for training
Input text is used solely for the immediate synthesis task. We do not aggregate or analyze your literary content to train our engines, keeping your intellectual property distinct from our training data.
No storage of generated audio
Generated audio files are delivered directly in the API response and are not persisted in our storage systems. This ensures that your final audio outputs remain transient and under your control.
Metadata kept for billing only
We retain request metadata, including timestamp, character count, and status, exclusively for accurate billing and usage tracking. This data is not used for analytics or marketing purposes.
Integration Simplicity
AudioLex Access requires no proprietary client libraries. By adjusting the base URL, existing OpenAI SDK implementations function immediately, reducing integration overhead for accessibility developers.
Authentication relies on standard bearer tokens. Each account maintains a single active key, ensuring consistent authorization across all synthesis requests without complex credential management.
OpenAI SDK compatibility
The API mirrors the standard speech endpoint structure. Developers can reuse established OpenAI client code, modifying only the base URL to point to the AudioLex Access service.
Base URL configuration
Set the base URL to https://api.6zzz.info/v1. This directs all POST requests for audio synthesis to the correct processing infrastructure without requiring custom network handlers.
Standard authentication
Include the Authorization header with your Bearer token (tts-...). A new key invalidates the previous one immediately, simplifying key rotation and security management.
Reliability and Limits
AudioLex Access provides predictable performance for dyslexia applications through defined rate limits and queue management. The API processes requests sequentially, ensuring consistent audio quality while managing server load. Developers can rely on clear boundaries for request volume and processing times during peak usage periods.
60 requests per minute
Each API key supports up to 60 requests per minute once the account is funded. New accounts without top-ups are limited to 10 requests per minute. This constraint ensures stable performance for indie developers building reading applications.
Queue processing under load
Requests are processed one after another. Under heavy load, a request may wait in a queue before synthesis begins. Typical generation speed is roughly real-time, producing about 25 characters of text per second of audio.
Retry logic for failures
Failed requests incur no charge. If synthesis fails, the cost is returned in full. Developers should implement retry logic to handle temporary processing delays or queue waits without incurring additional costs for unsuccessful attempts.
Use Case: Esperanza Rising
This section details the specific workflow for synthesizing "Esperanza Rising" for accessibility applications, leveraging the API's constraints to ensure consistent user experience. It addresses character limits, voice consistency, and chapter processing.
Chapter-by-chapter synthesis
Process each chapter as a discrete unit to maintain clear audio boundaries. This approach allows users to navigate chapters independently without downloading entire books. Each chapter request is processed sequentially.
- Submit one chapter per API call.
- Receive distinct audio files for each section.
- Simplify error handling per chapter.
Character limit management
Split text into segments of 4096 characters or fewer before sending requests. This ensures compliance with the input limit without truncating narrative flow. The API returns audio files that can be concatenated client-side for seamless playback.
- Limit: 4096 characters per request.
- Split text programmatically.
- Concatenate outputs for continuous listening.
Consistent voice across chapters
Use the same voice preset (e.g., Aria for narration) across all chapters. This maintains auditory continuity for listeners with dyslexia, reducing cognitive load. The voice remains identical regardless of the chapter length or language setting.
- Fixed voice ID per book.
- Consistent tone across sections.
- Supports 23 languages.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. WAV provides uncompressed, lossless audio, preserving every detail of the synthesized voice.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- IETF RFC 7845: Ogg Opus
- W3C WAI: making audio and video accessible
- Node.js documentation
- LibriTTS-R paper (arXiv 2305.18802)
- LibriTTS-R corpus (OpenSLR 141)
- Creative Commons Attribution 4.0
- curl documentation
- OpenAI API reference: create speech
- Python documentation
- ISO 639 language codes
- Tether (USDT)
