Frequently asked questions
Who runs the Kuwi editorial team?
An independent editorial team operates Kuwi Audio, focusing on reliable text-to-speech solutions for podcast publishers. The team manages the API infrastructure and ensures consistent voice quality for episodic content production without external corporate oversight.
Are the voices owned by Kuwi?
Kuwi provides access to voices based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. Kuwi does not claim ownership of the underlying voice performances but offers them as presets for your TTS API usage.
Where do the voices come from?
The eight voice presets are derived from licensed recordings in the LibriTTS-R dataset, which includes public-domain audiobook readings from LibriVox. These sources provide the foundational audio data used to generate the synthetic speech outputs delivered via the API.
Is Kuwi affiliated with OpenAI?
Kuwi Audio is not affiliated with OpenAI. While the API endpoint structure mimics the OpenAI speech endpoint for SDK compatibility, Kuwi is an independent storefront. We use our own keys, pricing, and infrastructure, distinct from OpenAI’s services.
How is my data handled?
Input text is used solely to produce the audio file and is not stored or used for training. Generated audio is returned in the response and not retained. We keep request metadata, such as time and character count, strictly for billing purposes.
Do I need to credit Kuwi?
Attribution requirements depend on the underlying license of the source recordings, typically CC BY 4.0 for LibriTTS-R. Kuwi does not mandate specific credit lines for API usage, but the CC BY 4.0 license may require attribution. Check the specific voice preset details for any additional attribution requirements from the original source.
Can I contact support via phone?
No, there is no phone number for support. Kuwi Audio provides support exclusively via email. We do not offer live chat, ensuring that all inquiries are documented and addressed by our independent team without the need for real-time voice interaction.
Is there a physical office?
Kuwi Audio operates as an independent entity without a listed physical office address. As a cloud-based API service, our infrastructure and editorial team function remotely, allowing us to maintain a lean structure focused on delivering reliable TTS services to podcast publishers.
Do you use my text for training?
No, your input text is used only to produce the immediate audio output. It is not stored for long-term use nor used to train our models. Generated audio is returned in the response and is not retained on our servers.
Are the voices royalty-free?
The voices are based on LibriTTS-R (CC BY 4.0) and LibriVox public-domain readings. While the API usage is royalty-free, you must adhere to the CC BY 4.0 license terms, which may require attribution depending on your distribution method.
How do I verify the license?
The voice presets are based on LibriTTS-R (CC BY 4.0) and LibriVox public-domain recordings. You can verify these licenses through the standard Creative Commons and LibriVox documentation. The API documentation provides details on the specific voice presets and their underlying data sources.
What is the team's background?
Kuwi Audio is an independent storefront for a text-to-speech API designed for podcast publishers. Our team focuses on providing reliable, scalable TTS solutions with predictable pricing. We prioritize consistent voice quality and efficient API integration for episodic content production workflows.
Do you have a privacy policy?
As an independent service, Kuwi Audio handles data according to our specific practices: input text is used only for generation, and metadata is kept for billing. We do not store generated audio or use text for training. Check our site for the full policy.
How can I provide feedback?
You can provide feedback via email, which is our sole support channel. We do not offer phone support or live chat. Your feedback helps us improve the API reliability and voice presets for podcast publishers using our text-to-speech services.
Our Editorial Mission
Kuwi Audio operates as an independent storefront, prioritizing clarity over marketing hype. We provide text-to-speech infrastructure designed specifically for podcast publishers who need consistent, broadcast-quality audio for episodic content.
Our approach relies on verifiable data rather than subjective claims. We focus on the technical realities of our API, ensuring that every feature and limitation is documented with precision for production workflows.
Expert Reviews
We evaluate our own service with the same rigor applied to third-party tools. Our reviews focus on practical metrics like character limits, voice presets, and processing speeds relevant to script-to-audio pipelines.
Every claim is backed by specific numbers, such as the $10 per million character rate or the 4096-character input limit, helping publishers plan their production budgets accurately.
- No subjective superlatives
- Focus on production workflows
Fact-Based Comparisons
Comparisons are drawn only from published facts and current API specifications. We do not invent competitors or exaggerate performance; we simply present the data as it exists.
This allows podcast producers to make informed decisions based on tangible constraints, such as the lack of SSML or the specific language support for 23 languages.
- Verified metrics only
- No invented statistics
Industry Focus
Our primary audience is podcast publishers who need to convert scripts into audio at scale. We understand the need for predictable pricing and consistent voice quality across episodes.
The API is structured to handle episodic content production, with clear rate limits and straightforward billing that aligns with the variable nature of publishing schedules.
- Designed for publishers
- Episodic content focus
Voice Licensing
We derive our voice presets from the LibriTTS-R dataset, which consists of readings sourced from LibriVox. These recordings are public domain audiobook narrations, providing a robust foundation for our eight distinct voice presets. This approach ensures consistent, high-quality synthetic speech without relying on proprietary, black-box models.
LibriTTS-R Source
Our eight voice presets—such as Aria, Nova, and Atlas—are trained on licensed recordings from the LibriTTS-R dataset. This dataset contains thousands of hours of speech, selected for clarity and broadcast readiness. By using these established recordings, we maintain a stable voice quality across all generated audio files.
- Source: LibriTTS-R dataset
- Origin: LibriVox public-domain readings
- Coverage: 8 distinct voice presets
CC BY 4.0 License
The underlying voice data is available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This allows us to legally integrate these voices into our API while respecting the original creators' attribution rights.
Public Domain Records
Many of the source recordings originate from public domain audiobooks. This historical foundation means our voices are not tied to exclusive talent contracts, ensuring long-term availability and predictable pricing. The synthetic nature of the output means no residual royalty payments are triggered by your publication volume.
Data Policy
Kuwi Audio processes your input text solely to generate the requested audio file. We do not use your script content to train our models or improve voice synthesis for other customers. The text is processed in real-time and discarded after synthesis, ensuring your editorial content remains private and exclusive to your production workflow.
Input Text Usage
Your input text is used exclusively for the immediate generation of the audio file. It is not retained for training purposes or used to refine the underlying synthesis engine. This ensures that your unpublished scripts and proprietary content remain confidential and are not absorbed into the model’s public knowledge base.
Audio Storage Policy
Generated audio files are returned directly in the API response and are not stored on our servers after delivery. You retain full control over the final audio assets, deciding whether to archive them locally or on your preferred podcast hosting platform. There is no hidden archival of your episodes by Kuwi.
Billing Metadata
We retain request metadata including timestamp, character count, voice used, and status code to ensure accurate billing and troubleshooting. This data is used solely to verify that charges correspond to actual synthesis requests and to maintain a clear audit trail for your accounting records. No content from your text is stored here.
Contact and Support
Kuwi Audio operates without a call center. We rely entirely on email for all customer inquiries, technical questions, and billing issues. This approach keeps our operations lean and focused on delivering a reliable API for podcast publishers.
There is no phone number listed on this site. If you require voice support, please use the contact form below or email us directly. We believe email provides a clear, searchable record of our technical exchanges.
Email Support Only
All support requests are handled via email. We do not offer live chat, phone, or ticketing systems. You can reach us at [email protected].
For API-specific issues, include your request ID and key prefix in the email body to help us diagnose the problem quickly.
No Phone Number
We do not maintain a public phone line. This reduces overhead and keeps our API pricing predictable for small podcast studios.
All communications are asynchronous, allowing our engineering team to respond with detailed technical answers rather than brief verbal summaries.
Response Times
During high-volume periods, responses may take slightly longer.

Company Structure
Kuwi Audio operates as an independent storefront for the same text-to-speech API available across the network. Each site runs autonomously, meaning your account and key are specific to playkuwi.com. This structure ensures that pricing, rate limits, and usage data remain isolated to this domain, providing clear accountability for podcast publishers managing episodic content production.
Independent Storefront
We function as a distinct commercial entity rather than a sub-brand of a larger tech conglomerate. Your API key, balance, and character usage are tracked exclusively within this storefront’s ecosystem. This independence allows us to maintain direct control over our infrastructure and billing systems without intermediary layers.
No Affiliation with Giants
Kuwi Audio is not affiliated with OpenAI, ElevenLabs, Google, Amazon, Microsoft, or any other major platform mentioned in industry comparisons. While our API endpoint shape mirrors OpenAI’s speech model for developer convenience, we are a separate company with distinct voice models, pricing, and infrastructure. We do not use their brand identity, and our voices are based on licensed recordings from LibriTTS-R and LibriVox, not their proprietary datasets.
Transparent Operations
Our operational model is straightforward: we charge $0.01 per 1,000 characters with no hidden fees or monthly subscriptions. All billing metadata is kept simple, and failed requests are free. We do not offer complex enterprise contracts or SLAs; instead, we provide a consistent, predictable API for podcasters who need reliable audio generation without negotiating custom terms.
- Single active key per account
- USDT (TRC-20) top-ups only
- No VAT invoices or subscriptions
Technology Stack
Kuwi Audio operates as an independent storefront for a cloud-based text-to-speech API. We do not own the underlying speech synthesis engine but provide a dedicated interface for podcast publishers to convert scripts into broadcast-quality audio at scale. Our infrastructure is built to handle episodic content production with predictable performance and consistent voice quality.
Cloud Infrastructure
Our service runs on a scalable cloud environment optimized for text-to-speech generation. Requests are processed sequentially, which ensures stability but may introduce queue delays during high load. We do not offer SLAs, uptime guarantees, or data-center choices, keeping operations simple and transparent for users who prioritize predictable pricing over enterprise-grade redundancy.
API Architecture
- Standard JSON request format
- OpenAI SDK compatible
- Single active key per account
Security Measures
Authentication is handled via a Bearer token in the Authorization header. Each account holds one active key; issuing a new key immediately invalidates the old one. We do not store generated audio files, and input text is used solely for synthesis. Request metadata is retained for billing purposes only.
- Bearer token authentication
- No audio storage
- Metadata kept for billing
Listen to the 8 voices
Each sample below was generated by this API from the same sentence.
23 languages
Any voice speaks any of these languages.
