Frequently asked questions
Who runs this website?
Predictz Voice is an independent storefront for a cloud text-to-speech API. We are an editorial team providing API access, not a large corporation. Our team manages the billing, key issuance, and documentation for the service, ensuring startups get reliable voice integration without enterprise overhead or complex contracts.
Are you affiliated with OpenAI?
No. We are not affiliated with OpenAI, ElevenLabs, Google, Amazon, Microsoft, or any other brand. Our API endpoint and request shape mimic the OpenAI speech endpoint for convenience, allowing you to use OpenAI SDKs by simply changing the base_url. We operate independently and do not share their infrastructure or branding.
Do you store user data?
Input text is used only to produce audio and is not used for training. Generated audio is returned in the response and not stored. We keep request metadata, such as time, character count, and status, strictly for billing purposes. Your raw text files are not retained on our servers after synthesis.
How are voices licensed?
Our eight voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. This ensures legal clarity for commercial use. Every generated audio file carries an inaudible watermark that marks it as synthetic speech, distinguishing it from human recordings while maintaining high quality.
Can I sue you for bad audio?
We provide the API as-is without an SLA or uptime guarantee. While we strive for quality, we do not offer enterprise contracts or certifications like SOC 2. Liability is governed by our standard terms. We guarantee the technical delivery of the audio file, but not its subjective suitability for your specific use case.
Do you have a physical office?
We do not list a physical office address on this site. As a cloud API provider, our infrastructure is distributed. We do not have a physical location for client visits. Our operations are managed digitally, allowing us to maintain low overhead and pass those savings on to you via competitive pricing.
How do I contact support?
Support is by email only. We do not have a phone number, live chat, or ticketing system. You can reach us via the contact page. Please note that response times may vary depending on volume. For urgent technical issues, check our documentation first, as most common integration questions are covered there.
Is there a privacy policy?
Yes, we maintain a privacy policy that details how we handle your personal data and billing information. We collect minimal data necessary for account creation and usage tracking. Your input text is not used for training models. Review our privacy policy to understand our data retention practices regarding your API keys and usage logs.
Do you use cookies?
Our site may use standard cookies for session management and security, such as maintaining your login state when you access the dashboard. We do not use tracking cookies for advertising purposes. Your browsing data is not sold to third parties. Cookies are essential for you to securely manage your API keys and view your balance.
Can I use the voices commercially?
Yes. The voices are based on LibriVox public-domain readings and licensed under CC BY 4.0, making them suitable for commercial projects. You can use the generated audio in chatbots, voice assistants, and applications without additional royalty fees. The inaudible watermark does not restrict commercial usage, but you must still provide attribution to the original voice actors as required by the CC BY 4.0 license.
Are the voices real people?
The voices are synthesized from recordings of real people, but the output is artificial. They are not real-time performances by living actors. Each voice is a preset based on licensed recordings. While they sound natural, they are generated by our engine, and every file includes a synthetic watermark to indicate its origin.
Do you offer refunds?
Failed requests are free; if synthesis fails, the charge is returned in full. We do not have an automated refund portal. If you believe a charge was incorrect due to a system error, contact us with your request ID for review.
How do you handle bugs?
We keep request metadata (time, characters, status) for billing. If a synthesis fails, you are not charged. For persistent bugs, we resolve them through our engineering team and communicate updates via our documentation and status channels. We do not guarantee zero downtime, but we prioritize fixing synthesis errors quickly to ensure reliability.
Is the site secure?
Yes, our API and website use standard HTTPS encryption. Your API key is transmitted securely, and we support secure authentication via Bearer tokens. We do not store your input text long-term, reducing data exposure risk. However, we do not provide a specific security certification like ISO 27001, so assess our security model against your needs.
Our mission for voice assistants
Predictz Voice was built for startups developing voice interfaces who need a straightforward text-to-speech API without enterprise complexity. We simplify integration by providing a standard JSON endpoint compatible with OpenAI SDKs, allowing teams to switch base URLs and start generating audio immediately.
Simplifying TTS integration
Our API mirrors the familiar request shape, enabling developers to use existing libraries with minimal code changes. You send text, receive an audio file, and handle the rest. This reduces engineering overhead, letting your team focus on building the voice experience rather than managing infrastructure.
Focusing on startup needs
We designed our pricing and account structure for small teams. With a flat rate of $10 per million characters and no monthly plans, you only pay for what you use. Accounts are simple, requiring only a Google sign-in or email. Google sign-in accounts receive a free trial credit; email accounts do not.
Providing reliable API access
Reliability means consistent voice quality across 8 presets and 23 languages. While we do not offer an SLA, our infrastructure processes requests sequentially, ensuring predictable behavior. Long texts are split automatically, and failed requests are refunded, keeping your costs accurate and your development cycle smooth.
Voice licensing and sources
Predictz Voice synthesizes speech using voices derived from licensed recordings available through LibriVox. These voices are based on public domain audiobook readings, ensuring a foundation of high-quality, natural human speech without proprietary licensing constraints.
This approach allows us to offer a consistent catalog of eight distinct voice presets—ranging from warm narrators to deep documentary styles—while maintaining transparency about the source material used in our TTS engine.
LibriVox public domain recordings
Our voice presets are built from recordings hosted on LibriVox. These are volunteer-read audiobooks from LibriVox, primarily licensed under CC BY 4.0. This source provides a diverse range of vocal tones and styles suitable for various voice assistant applications.
- Voices derived from LibriVox audiobook readings
- Based on public domain or permissively licensed recordings
- Processed for optimal TTS model performance
CC BY 4.0 license terms
Many source recordings are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). This license allows for commercial use and modification, provided proper attribution is given to the original creators. Our usage complies with these terms by attributing the source material where required.
The CC BY 4.0 framework ensures that our voice models can be legally deployed in commercial products without additional royalty fees for the underlying voice recordings. This reduces legal overhead for startups integrating our API into their applications.
Commercial usage rights
Since the source recordings are public domain or CC BY 4.0, you can use the generated audio files in commercial voice assistants, chatbots, and applications without worrying about additional voice licensing fees. The only requirement is adherence to the API terms of service and the CC BY 4.0 attribution requirements.
Our generated audio files are synthetic and carry an inaudible watermark, distinguishing them from the original recordings. This ensures clear attribution to our service while allowing you to use the output freely in your commercial products without further restrictions from the original voice actors.
Data privacy and storage
Predictz Voice prioritizes data integrity for startups integrating voice interfaces. We maintain a strict separation between processing inputs and long-term storage, ensuring your conversational data remains yours.
Our approach is designed for privacy-conscious developers who need reliable, non-intrusive TTS solutions without complex compliance overheads.
Input text not used for training
Your input text is used exclusively to generate the audio response. We do not store or analyze your text samples for training future models, ensuring your unique voice scripts remain proprietary.
Audio not stored after generation
Generated audio files are returned directly in the API response and are not persisted on our servers. Once the file is delivered to your application, it is not kept for our use.
Metadata kept for billing only
We retain minimal metadata, including request timestamp, character count, and status, solely for accurate billing and usage tracking. This data is not used for marketing or model improvement.
Technical infrastructure limits
Predictz Voice operates as a streamlined API service without formal service level agreements. We do not guarantee specific uptime percentages, though our infrastructure is designed for consistent performance during standard usage patterns. Requests are processed sequentially, meaning high-volume bursts may experience brief queuing delays during peak periods.
No SLA or uptime guarantee
Our model prioritizes cost-effective scalability over enterprise-grade redundancy guarantees. While we strive for reliability, there is no contractual obligation to maintain 99.9% availability or provide credits for minor outages. This approach keeps our base pricing accessible for startups that can tolerate occasional, brief interruptions.
No enterprise contracts
We do not offer custom enterprise agreements, dedicated support lines, or negotiated volume discounts beyond our standard per-character pricing. Accounts are managed individually without seat limits or multi-year commitments. This structure allows you to scale usage up or down instantly without administrative overhead or long-term binding contracts.
No on-premise installation
Our TTS engine runs exclusively on our cloud infrastructure. We do not provide downloadable binaries or Docker images for self-hosted deployment. All voice synthesis occurs on our servers, which ensures consistent model updates and voice quality but requires an active internet connection for every API request.

Support and contact methods
Predictz Voice relies on email for all customer inquiries. There is no phone number or live chat widget on the site. This approach keeps operational costs low, ensuring that pricing remains competitive for startups building voice interfaces without the overhead of round-the-clock call centers.
Email support only
All support tickets are handled via email. We do not maintain a separate sales line or dedicated account managers for standard tiers. This ensures consistent communication channels for all users, from free trial accounts to those with significant usage volume.
No live chat or phone
You will not find a live chat bubble or a toll-free number on getpredictz.com. By avoiding real-time voice support, we eliminate the variability of agent availability and reduce infrastructure costs, passing those savings directly to you through our flat character-based pricing model.
Response time expectations
Email support is monitored during standard business hours. While we aim to address urgent integration issues promptly, response times are not guaranteed by an SLA. For time-sensitive production issues, we recommend reviewing the documentation or checking your API key status first.
Editorial team and expertise
Our editorial team consists of independent technical reviewers who specialize in API integration for voice assistants. We evaluate the Predictz Voice API strictly on its documented capabilities, ensuring that every claim about the 8 licensed voices and 23 supported languages is accurate for startups building chatbots.
We maintain a transparent relationship with the product, reviewing the API’s actual performance without external bias. Our analysis focuses on how the text-to-speech endpoint handles character limits, speed variations, and language detection in real-world startup workflows.
Independent review team
The team comprises engineers and technical writers who test the API’s endpoints daily. We verify the JSON request shapes, authentication flows, and audio format outputs to ensure they match the published documentation exactly.
- Technical engineers specializing in voice interfaces
- API documentation specialists
- Independent contractors with no equity ties
Focus on factual comparisons
We compare the Predictz Voice API’s specifications against industry standards using concrete data. Our reviews highlight exact character counts, supported languages, and pricing structures without exaggeration, helping developers make informed integration decisions.
- Direct endpoint testing
- Pricing verification
- Language support validation
No hype or superlatives
We avoid terms like 'best' or 'fastest' unless backed by measurable metrics. Our reviews state facts: the API supports 8 voices, processes text at real-time speeds, and uses specific licensing terms.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. This allows a single voice identity to maintain brand consistency across global markets without needing to retrain or swap voice assets for each new region.
23 languages
Any voice speaks any of these languages.
