Frequently asked questions
Are you affiliated with OpenAI?
No, we are not affiliated with OpenAI. Our API endpoint and request shape mimic the OpenAI speech endpoint for compatibility, but we are an independent storefront. You use the official OpenAI SDKs by simply changing the base_url to our domain. We do not use OpenAI’s infrastructure or brand.
Where do the voices come from?
Our eight voice presets are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. Each generated audio file includes an inaudible watermark to mark it as synthetic speech. We do not offer custom voice cloning or upload your own voice.
Do you sell user data?
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. We keep request metadata (time, characters, status) solely for billing purposes. We do not sell or share your data.
How is the audio generated?
We send your text to our cloud API, which synthesizes the audio using licensed voice models. For texts over 4096 characters, we split them into parts and join the resulting audio files. The process takes roughly real-time duration, with short phrases taking a few seconds to generate.
Is the API stable?
We do not provide 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. The rate limit is 60 requests per minute per key to ensure consistent performance for your language learning materials.
Do you have a physical address?
. We are a cloud-based API service. Our support is handled entirely via email. .
Can I contact you by phone?
No, we do not have a phone number or live chat. All support is provided via email only. You can reach us through the contact page or by emailing the support address provided in your account dashboard. We aim to respond to all inquiries promptly.
Do you offer custom voices?
No, we do not offer custom voice cloning or voice design. We provide eight preset voices (Aria, Nova, Luna, Sage, Felix, Milo, Atlas, Orion) based on licensed recordings. You cannot upload your own voice or create a unique voice model for your specific language learning content.
Who We Are
ArionVoice is an independent team building a practical text-to-speech API for educators and methodologists. We focus on delivering consistent, high-quality audio for language learning materials without the overhead of large corporate structures.
Our goal is to help teachers generate clear voice-overs for videos, podcasts, and classroom resources efficiently. We provide a straightforward tool that integrates easily into your existing workflow, letting you concentrate on content rather than infrastructure.
Independent Team
We are a small, dedicated group focused on reliability. Unlike large tech firms, we make direct decisions about our API’s behavior and pricing, ensuring it remains stable and predictable for long-term educational projects.
Focus on Education
Our interface and documentation are designed for clarity. We support 23 languages and clear audio output suitable for classroom use, ensuring students hear consistent speech patterns.
No Corporate Baggage
There are no complex enterprise contracts or hidden fees. You pay only for what you use, with a simple balance system that does not expire. This transparency makes it easy to budget for course production.
- Simple pricing: $10 per million characters
- No monthly subscriptions or seats
- Transparent billing metadata only
Our Voices and Licensing
ArionVoice delivers eight distinct voice presets—such as Aria, Nova, and Atlas—designed for clear educational narration. These voices are based on licensed recordings from the LibriTTS-R dataset, ensuring consistent audio quality suitable for language learning materials.
LibriTTS-R Source
Our voice models are derived from the LibriTTS-R dataset, which contains high-quality audio readings. This foundation allows us to offer eight distinct presets, including warm narrators and calm instructors, optimized for clarity in classroom resources.
CC BY 4.0 License
The source recordings are licensed under Creative Commons Attribution 4.0 (CC BY 4.0). This license permits commercial use and modification, providing a clear legal framework for integrating these voices into your educational content without complex royalty tracking.
LibriVox Public Domain
Many of the source recordings originate from LibriVox, a platform for public-domain audiobook readings. By leveraging these existing public domain and CC-licensed assets, we maintain a straightforward licensing model that avoids the complexities of custom voice licensing.
Data Policy
When you send text to our API, the input is processed solely to generate the audio file and is immediately discarded. We do not use your classroom materials, lesson plans, or student readings to train or improve our models. Your content remains yours.
No Training Data
Unlike many commercial TTS services, ArionVoice does not ingest customer text for algorithmic training. The voices are pre-trained on public domain recordings, and your specific inputs are treated as transient data. This ensures that your unique educational content is never used to shape future voice outputs.
No Audio Storage
Generated audio files are returned directly in the API response and are not stored on our servers after delivery. Once you receive the audio file, our system no longer holds a copy. This minimizes data retention risks and keeps your generated voice-overs strictly under your control.
Billing Metadata Only
We retain minimal metadata strictly for billing accuracy: the timestamp, character count, and request status (success or failure). This data is used solely to calculate your usage charges and verify that failed requests are refunded. It does not include the text content or the audio file itself.
Technical Infrastructure
ArionVoice runs as a cloud-based API, eliminating the need for local infrastructure or complex server management. You simply send text via HTTP POST requests and receive audio files directly, making it easy to integrate with existing lesson planning tools or learning management systems without hiring DevOps specialists.
Cloud-Based API
Access the service from anywhere with an internet connection. The API handles the heavy lifting of speech synthesis in the cloud, ensuring consistent performance regardless of your local hardware. This simplicity allows teachers and methodologists to focus on content creation rather than technical maintenance.
- RESTful API with standard JSON requests
- Compatible with OpenAI SDKs by changing the base_url
- No local software installation required
Simple Architecture
The system processes requests sequentially, which means under heavy load, synthesis may queue but remains reliable. Text is split into 4096-character chunks if necessary, and audio files are joined automatically. This straightforward design reduces potential points of failure, ensuring predictable output for your educational materials.
No Complex SLAs
We do not offer enterprise-grade Service Level Agreements with uptime guarantees or financial penalties for downtime. Our architecture is designed for cost-effective, reliable access rather than guaranteed 99.99% availability. This keeps prices low and predictable, suitable for individual educators and small teams who prioritize value over corporate guarantees.
- No uptime guarantees
- No penalty clauses
- Transparent, straightforward service terms

Support and Contact
ArionVoice operates as a lean, independent API provider without the overhead of a large corporate support department. We prioritize building a reliable product that teachers and methodologists can integrate directly into their language learning workflows, ensuring consistent audio generation for educational materials without unnecessary complexity.
Our communication structure is designed to be transparent and efficient. We do not maintain a dedicated phone line or a live chat widget that might promise immediate answers we cannot guarantee. Instead, all inquiries are handled through a single, reliable email channel to ensure no question is lost in a queue.
Email Only
All support requests, technical questions, and billing inquiries are directed to our support email. This single point of contact ensures that your issue is logged and addressed by a team member who understands the specific constraints of our TTS API.
No Live Chat
We do not offer live chat or a phone number. This eliminates the expectation of real-time availability and allows us to provide thoughtful, documented responses to your technical or account-related questions without the pressure of instant resolution.
Response Time
While we do not guarantee a specific SLA, . Given our small team size, complex technical queries regarding voice selection or API integration may take slightly longer to resolve thoroughly.
- No guaranteed SLA or uptime contract
- No phone support available
- Email is the only contact method
Honest Limitations
ArionVoice delivers complete audio files via a standard HTTP POST request. We do not offer real-time streaming or WebSocket connections, so you must wait for the full synthesis to complete before playing the audio. This design keeps our infrastructure simple and reliable for batch processing educational materials.
No Streaming
The API returns the final audio file (MP3, WAV, etc.) once generation is finished. There is no partial audio stream, meaning you cannot play the audio while it is still being generated. For long texts, we split the input into 4096-character chunks and join the resulting files, which still requires waiting for each chunk to complete.
- Standard HTTP response with full audio file
- No WebSocket or streaming endpoints
- Ideal for batch processing, not live interaction
No SSML
We do not support Speech Synthesis Markup Language (SSML). You cannot use tags to control pauses, pitch, or emphasis manually. The voice engine uses its internal pacing based on the text structure. If you need precise control over pronunciation or timing, you must adjust the input text or use the speed parameter.
- Plain text input only
- No pause, pitch, or prosody tags
- Speed control available via parameter
No Voice Cloning
We do not offer custom voice cloning or uploading your own voice samples. The eight available voices (Aria, Nova, Luna, Sage, Felix, Milo, Atlas, Orion) are fixed presets based on licensed recordings. You cannot create a unique voice profile for your brand or students. The voices are generated from public-domain audiobook readings and licensed under CC BY 4.0.
- 8 fixed preset voices
- No custom voice upload
- No cloning service available
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This voice adds a layer of professionalism to advanced course materials.
23 languages
Any voice speaks any of these languages.
