Frequently asked questions
Who runs this site?
VoiceCompliance operates this independent storefront for the text-to-speech API. We provide the infrastructure for product teams managing high-volume video voice-overs. Our platform is distinct from other providers, offering a reliable API endpoint for generating accessible audio files directly from text inputs.
Are the voices licensed?
Yes. Our eight voice presets are based on licensed recordings from LibriTTS-R, which is licensed under CC BY 4.0. These sources include LibriVox public-domain audiobook readings. This ensures that the synthetic voices used in your generated audio files are properly licensed for commercial use.
Do you use my data for training?
No. Input text is used exclusively to produce the requested audio file and is not utilized for training models. Generated audio is returned in the response and is not stored. We retain request metadata, such as time, character count, and status, solely for billing purposes.
How do I contact support?
Support is available via email only. We do not offer phone support, live chat, or on-site assistance. For technical issues or billing inquiries, please send an email to our support team. Response times may vary depending on the volume of requests we are processing.
Is the site affiliated with OpenAI?
No. VoiceCompliance is an independent storefront and is not affiliated with OpenAI, ElevenLabs, Google, Amazon, Microsoft, or any other brand mentioned in comparisons. Our API endpoint shape matches the OpenAI speech endpoint for compatibility, but we operate independently and do not share ownership or branding.
Do you offer refunds?
Failed requests are free; if synthesis fails, the charge is returned in full. We do not offer manual refunds for successful requests. There are no monthly plans or subscriptions, so you only pay for what you use. Balance does not expire, and top-ups are via USDT on the TRON network (TRC-20) only.
How are voices selected?
We offer eight presets: Aria, Nova, Luna, Sage, Felix, Milo, Atlas, and Orion. These are based on licensed recordings from LibriTTS-R and LibriVox. Each voice has a distinct tone, such as warm, bright, or deep, allowing you to select the most appropriate voice for your accessibility standards.
Is the team remote?
We do not have a physical office or a public team structure. Support is handled via email only, with no phone number or live chat available. The team manages the API infrastructure and billing systems remotely, ensuring consistent uptime and response to technical issues via email.
Do you have an office?
No. We do not maintain a physical office or on-premise installation option. Our entire operation, including API servers and billing systems, is cloud-based. This allows us to provide a scalable, reliable text-to-speech API without the overhead of physical locations or local data centers.
Are the voices real people?
The voices are synthetic, generated from licensed recordings (LibriTTS-R) of real people. Each generated file carries an inaudible watermark marking it as synthetic speech. While the source material comes from real speakers, the output is AI-generated audio, not a direct recording of a human speaking your text.
Can I sue for copyright?
Our voices are based on LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. We do not offer custom voice cloning, so you cannot claim rights to a specific person’s voice unless you commissioned a custom recording, which we do not provide. Consult legal counsel for specific use cases.
Do you delete data?
Input text is used only to produce audio and is not stored. Generated audio is returned in the response and not stored. We keep request metadata (time, characters, status) for billing purposes. We do not retain the actual text or audio files after the request is completed.
Is the API open source?
No. We do not provide an official SDK, nor is the API code open source. You can use the OpenAI SDK by changing the base URL or use plain HTTP requests. The API endpoint is proprietary, and we do not publish our source code or server implementations.
Who owns the audio?
You own the audio files generated from your text input. We provide the synthetic voice engine and infrastructure. The voices themselves are based on licensed recordings, but the specific audio file you generate is yours to use. We do not store the audio, so you retain full control.
Our Editorial Team
VoiceCompliance operates as an independent editorial entity, curating synthetic voice assets specifically for high-volume video production teams. We do not manufacture the underlying audio models but strictly select and validate eight distinct voice presets to ensure consistent output for accessibility and dyslexia-friendly applications.
Independent Curation
Our team manages the selection of voices based on clarity and acoustic stability rather than celebrity appeal. We verify that each voice delivers consistent pronunciation across 23 supported languages, though the language must be explicitly set for non-English Latin-script text to ensure correct pronunciation., ensuring reliable comprehension for diverse audiences without requiring manual post-production cleanup.
- Curated selection of 8 voice presets
- Focus on acoustic clarity and stability
- No celebrity or influencer voice mimicry
Technical Expertise
We maintain rigorous control over the TTS API infrastructure, ensuring compatibility with standard OpenAI SDK workflows. Our technical approach prioritizes deterministic output, allowing product teams to integrate speech generation into their existing CI/CD pipelines for automated video voice-overs.
- API-first architecture
- OpenAI SDK compatibility
- Deterministic audio generation
Accessibility Focus
Our editorial standards prioritize voices that support readability for users with dyslexia. We select audio profiles that avoid excessive emotional modulation, providing neutral, clear narration that enhances content retention and meets strict accessibility guidelines for digital media.
- Dyslexia-friendly audio profiles
- Neutral tonal delivery
- Enhanced content retention
Voice Licensing
Our eight voice presets are derived from the LibriTTS-R dataset, sourced from LibriVox public-domain audiobook readings. This foundation ensures that our synthetic voices are built upon licensed recordings (CC BY 4.0) that are freely available for commercial use.
Each generated audio file carries an inaudible watermark to clearly mark it as synthetic speech, providing transparency for accessibility standards and compliance audits.
LibriTTS-R Source
The voice models are trained on the LibriTTS-R dataset, which consists of public-domain audiobook readings from LibriVox. This source provides a diverse range of natural speech patterns suitable for high-volume video voice-overs.
CC BY 4.0 License
The underlying recordings are licensed under CC BY 4.0, allowing for unrestricted commercial use. This simplifies licensing for product teams managing large-scale audio content without complex royalty calculations.
Public Domain Basis
Since the source material is based on recordings licensed under CC BY 4.0, the resulting synthetic voices inherit this clear legal standing. This reduces the risk of copyright claims for your video productions.
Data Privacy Policy
VoiceCompliance processes input text solely for the immediate generation of audio files. Your source content is not ingested into our training datasets, ensuring that proprietary scripts and sensitive narratives remain distinct from model development. This separation guarantees that your voice-over materials do not influence future voice synthesis outcomes or become part of the underlying engine’s knowledge base.
No Training on Input
Text submitted via the API is used exclusively to produce the requested audio file. We do not retain copies of your input text beyond the immediate processing window, nor do we use it to train or refine our voice models. This ensures that your specific content remains confidential and does not contribute to the general training corpus.
No Audio Storage
Generated audio files are returned directly in the API response and are not stored on our servers after delivery. Once the audio is transmitted to your client application, the file is discarded. This minimizes data exposure and ensures that your voice-over assets are transient unless you choose to save them in your own infrastructure.
Billing Metadata Only
We retain minimal metadata strictly for billing accuracy and request auditing. This includes the timestamp, character count, voice selected, and request status. This data is used solely to verify charges against your account balance and to troubleshoot technical issues. No content from the audio or text is analyzed for business intelligence or advertising purposes.
- Request timestamp and status
- Character count per request
- Selected voice and model
- Duration of processing
Technical Infrastructure
VoiceCompliance operates as a cloud-based text-to-speech API, eliminating the need for on-premise infrastructure or complex hardware setups. Our architecture is designed to handle high-volume requests for product teams managing video voice-overs, ensuring consistent delivery of synthetic audio files via standard HTTP protocols.
Cloud-Based API
The service is entirely cloud-hosted, allowing seamless integration into existing CI/CD pipelines. Requests are processed on our servers, with input text up to 4096 characters converted to audio and returned in the response; longer texts must be split by the client. Metadata such as timestamps and character counts are retained solely for billing purposes, while the generated audio itself is not stored.
OpenAI SDK Compatible
We utilize the same request shape as the OpenAI speech endpoint. By changing the base_url to https://api.368g.info/v1, teams can use official OpenAI SDKs directly. This ensures compatibility with existing codebases without requiring custom wrappers or migration efforts.
- Same JSON body structure as OpenAI
- Compatible with official OpenAI SDKs
- No custom SDK required
Standard HTTP Protocol
All interactions occur over standard HTTP. Authentication is handled via a Bearer token in the header. The API supports POST for speech generation and GET for retrieving voice lists and model information. This simplicity allows for easy integration with any language or framework that supports HTTP requests.

Support and Contact
VoiceCompliance operates with a lean, asynchronous support model designed for product teams who prefer documented resolution over real-time interruption. When technical or billing inquiries arise, our team reviews and responds via email. This ensures that every communication is captured, traceable, and available for future reference within your internal audit logs.
Email Support Only
All support requests are handled exclusively through email. This channel allows our engineers to attach relevant logs, code snippets, or configuration details directly to the response, ensuring clarity and reducing the back-and-forth typical of live conversations. For most API integration questions, you can expect a response via email.
No Live Chat
We do not maintain a live chat widget on our site. This decision reduces latency overhead and allows our support staff to focus on deeper technical analysis rather than managing multiple simultaneous short queries. If your issue requires immediate context switching, email remains the most efficient and reliable method of communication.
No Phone Number
There is no dedicated phone line for VoiceCompliance. We avoid phone support to maintain a consistent, written record of all support interactions, which is critical for compliance and accessibility workflows. If you require a verbal explanation, you may request a scheduled email exchange with detailed notes or a recorded video response.
Limitations and Honesty
VoiceCompliance delivers reliable synthetic speech through a standard cloud API, but our architecture prioritizes consistency over real-time interactivity. We do not offer streaming or WebSocket endpoints for partial audio delivery, meaning clients must await full synthesis before playback. Additionally, we do not provide custom voice cloning, allowing users to select from eight licensed presets rather than uploading unique vocal profiles.
Our service operates without a formal Service Level Agreement (SLA) or uptime guarantee. While we maintain stable infrastructure, we do not promise specific availability percentages or financial penalties for downtime. This structure allows us to allocate resources efficiently without the overhead of enterprise-grade contractual obligations, keeping costs predictable for high-volume product teams.
No Streaming Support
Our API returns complete audio files rather than streaming chunks. This design ensures that each generation is fully processed before delivery, reducing the likelihood of partial corruption during transmission. Clients requiring immediate playback must implement buffering logic to handle the latency between request submission and file receipt, which varies based on text length and server queue status.
No Custom Cloning
We do not offer custom voice cloning or the ability to upload and train on personal audio recordings. Users must choose from eight pre-defined presets (Aria, Nova, Luna, Sage, Felix, Milo, Atlas, Orion) based on licensed LibriTTS-R recordings. This limitation simplifies licensing compliance, as all generated audio carries an inaudible synthetic watermark and falls under the same CC BY 4.0 framework.
No SLA Guarantee
VoiceCompliance does not provide a formal Service Level Agreement (SLA) or uptime guarantee. We do not promise specific availability percentages or offer financial credits for service interruptions. This approach allows us to maintain cost efficiency for high-volume video voice-over workflows without the administrative overhead of enterprise contracts, though we strive for consistent performance across all regions.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. His natural pacing aids comprehension, particularly for accessibility-focused materials where clear, approachable speech is prioritized over formal narration.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- openai-node on GitHub
- LibriTTS-R corpus (OpenSLR 141)
- OpenAI API reference: create speech
- LibriTTS-R paper (arXiv 2305.18802)
- TRON developer hub: TRC-20
- IETF RFC 7845: Ogg Opus
- FFmpeg documentation
- Creative Commons Attribution 4.0
- W3C WAI: making audio and video accessible
- Node.js documentation
- ISO 639 language codes
