Frequently asked questions
Who runs this site?
This site is an independent storefront for the AudioLex Access text-to-speech API. We provide the synthesis infrastructure so developers can focus on user experience and compliance. We are not affiliated with OpenAI, ElevenLabs, Google, Amazon, Microsoft, or any other brand mentioned in comparisons.
Are the voices real people?
The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. Every generated file carries an inaudible watermark marking it as synthetic speech. We do not offer custom voice cloning or upload your own voice.
Is my data used for training?
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. Request metadata (time, characters, status) is kept for billing purposes only.
How do I contact support?
Support is available by email only. We do not offer phone support, live chat, or on-site contact options. Please use the email address provided on our contact page for any inquiries regarding your account or API usage.
Is there a phone number?
No, we do not have a phone number. All support communication is handled via email. This allows us to maintain a direct and documented line of communication with our developer clients without the overhead of voice support infrastructure.
Do you have an SLA?
We do not offer an SLA or uptime guarantee. Requests are processed sequentially, so under load a request may wait in a queue. Typical speed is roughly real time, but performance depends on current server load and queue depth.
Are the voices licensed?
Yes, our voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. This ensures that the synthetic speech outputs are generated from legally sourced source material for commercial and personal use.
Is the data stored after generation?
Generated audio is returned in the response and not stored by us. Input text is used only to produce the audio and is not used for training. We retain request metadata (time, characters, status) solely for billing accuracy.
Do you offer refunds?
Failed requests are free, and the charge is returned in full if synthesis fails. For other cases, we do not issue manual refunds. Your balance does not expire, so unused credits remain available for future API requests indefinitely.
Is the site affiliated with OpenAI?
No, we are not affiliated with OpenAI, ElevenLabs, Google, Amazon, Microsoft, or any other brand. We simply use the same request shape as the OpenAI speech endpoint, allowing you to use the official OpenAI SDKs by changing the base URL.
Can I request a feature?
We do not have a public feature request board. As an independent storefront, we prioritize core API stability. We do not offer streaming, SSML, or custom voice cloning. Please contact us via email to discuss specific needs for your application.
How is the balance managed?
Top-ups are made via USDT on the TRON network (TRC-20), minimum $5, credited automatically within minutes. No cards, PayPal, or subscriptions are used. Balance does not expire. Failed requests are free, and charges are returned in full upon failure.
Is the API GDPR compliant?
We do not hold a GDPR certificate. Input text is used only to produce the audio and is not stored or used for training. We retain request metadata for billing purposes. Compliance depends on your own data processing agreements.
Do you have a privacy policy?
Input text is not stored after generation and is not used for training. Generated audio is returned in the response and not stored by our servers.
Editorial Team and Origin
AudioLex Access is developed by an independent team focused on building reliable cloud infrastructure for text-to-speech applications. We operate as a standalone entity, distinct from larger technology corporations, ensuring our development priorities align directly with the needs of accessibility and dyslexia-focused reading software.
Independent development team
Our engineering and product teams manage the entire synthesis pipeline, from API endpoint maintenance to voice preset selection. This vertical integration allows us to maintain strict control over request processing and queue management without reliance on third-party hyperscalers for core audio generation.
No affiliation with major brands
We are not affiliated with OpenAI, ElevenLabs, Google, Amazon, or Microsoft. While our API endpoint structure mimics the OpenAI speech endpoint for SDK compatibility, we operate independently. Our brand, AudioLex Access, and our infrastructure are entirely separate from any major tech brand mentioned in comparative contexts.
Focus on accessibility tools
Our primary use case involves serving indie developers building reading applications for users with dyslexia. By providing a straightforward API that handles character limits and audio joining, we allow developers to concentrate on user experience and compliance rather than managing synthesis infrastructure.
Voice Licensing and Sources
AudioLex Access synthesizes voice data from licensed recordings and public-domain audiobook readings. This foundation ensures consistent, high-quality output for accessibility applications without requiring complex voice design or custom cloning.
LibriTTS-R dataset usage
Our eight voice presets are derived from the LibriTTS-R dataset. This provides a standardized base for all generated audio, ensuring that every request produces consistent synthetic speech suitable for reading applications.
CC BY 4.0 licensing
The underlying recordings utilize the CC BY 4.0 license. This allows us to process the audio data efficiently while maintaining clear attribution standards for the source material used in our API infrastructure.
LibriVox public domain sources
The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and from LibriVox public-domain audiobook readings. Any voice can speak any of these 23 languages.
Data Privacy Policy
AudioLex Access prioritizes data integrity for developers building accessibility tools. Our architecture ensures that input text is used exclusively for the immediate synthesis of audio files and is not retained for model training or secondary analysis.
Text not used for training
We do not store input text after audio generation. Your content remains transient, ensuring that proprietary narratives or user data are not repurposed for improving our engine.
No audio storage
Generated audio files are returned directly in the API response and are not stored on our servers. Once the download is complete, the audio file is effectively ephemeral.
Metadata retention for billing
We retain request metadata, including timestamp, character count, and status, solely for accurate billing and service monitoring. This data does not contain the actual text or audio content.
Technical Infrastructure
AudioLex Access operates as a cloud-based text-to-speech API, designed for integration into accessibility and dyslexia-focused reading applications. The service handles all synthesis infrastructure, allowing developers to focus on user experience and compliance without managing hardware.
The request shape is the same as the OpenAI speech endpoint, so the official OpenAI SDKs work by changing base_url to https://api.6zzz.info/v1. Requests are processed sequentially, with typical generation speeds aligning closely with real-time playback for longer texts.
Cloud-based deployment
The service runs entirely in the cloud, utilizing a distributed processing pipeline. Text is submitted via JSON POST requests to https://api.6zzz.info/v1/audio/speech, and audio files are returned directly in the response. This model eliminates the need for local audio engine dependencies.
No on-premise options
There is no on-premise installation option available. All computation occurs on our servers, meaning your application requires an internet connection to generate speech. This ensures consistent voice quality and immediate access to updates without manual version management on client hardware.
Standard API endpoints
The API follows a standard RESTful structure. Voice metadata is accessible via GET /v1/voices, and model information via GET /v1/models. Authentication is handled via a Bearer token in the request header, compatible with standard HTTP clients and the official OpenAI SDKs.
- POST /v1/audio/speech for synthesis
- GET /v1/voices for voice metadata
- GET /v1/models for model details
- Bearer token authentication

Support and Contact
AudioLex Access operates as a lean, independent infrastructure provider for text-to-speech synthesis. We prioritize technical reliability over extensive customer service layers, ensuring that developers building accessibility tools receive direct, unambiguous communication regarding their API usage and account status.
Email-only support
All inquiries are handled via email. This channel allows our engineering team to review technical details, API keys, and billing metadata thoroughly before responding. We do not maintain dedicated support queues for individual tickets; instead, we process requests in the order they are received.
- Include your API key and error logs for faster resolution
- No ticket number system; reference your email in follow-ups
No live chat
We do not offer live chat, phone support, or instant messaging widgets on this site. This absence is intentional, keeping operational costs low and allowing us to focus on API stability. If you require immediate assistance during a critical integration, email remains the most reliable contact method.
Response time expectations
Response times vary based on current request volume. During peak usage periods, replies may take several hours. We aim to address billing issues and technical errors promptly, but do not guarantee a specific response window. For urgent matters, consider checking our documentation first, as most common integration issues are resolved there.
Limitations and Honesty
AudioLex Access operates without a formal Service Level Agreement. We do not guarantee specific uptime percentages, nor do we maintain enterprise contracts for priority support. Our infrastructure is designed for reliability, but we do not promise a fixed availability target.
No SLA guarantee
We provide a standard cloud API without a contractual Service Level Agreement. This means there is no guaranteed uptime percentage or financial penalty for service interruptions. Our rate limits and processing times are best-effort, particularly during high-load periods when requests may queue.
- No contractual uptime percentage
- No financial penalty for downtime
- Best-effort processing under load
No uptime promise
We do not guarantee continuous service availability. While we aim for stable operation, unexpected infrastructure issues may occur without prior notice. Developers should implement retry logic in their applications to handle transient failures gracefully.
- No guaranteed 99.9% or similar metric
- No prior notice for maintenance
- Client-side retry recommended
No certifications claimed
We do not hold external compliance certifications such as SOC 2 or HIPAA. Our data policy is transparent: input text is not used for training, and generated audio is not stored after delivery. We rely on clear contractual terms rather than third-party audits to define our service boundaries.
- No SOC 2 certification
- No HIPAA compliance
- No GDPR certificate
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
- openai-node on GitHub
- FFmpeg documentation
- ISO 639 language codes
- OpenAI API reference: create speech
- TRON developer hub: TRC-20
- W3C WAI: making audio and video accessible
- Creative Commons Attribution 4.0
- LibriTTS-R paper (arXiv 2305.18802)
- Python documentation
- IETF RFC 7845: Ogg Opus
- MDN: Web Speech API
