About SynthLogic and Our Text-to-Speech Infrastructure

By SynthLogic editorial team · Updated 2026-11-18

This page outlines the operational and editorial standards of SynthLogic, an independent storefront for the SynthLogic text-to-speech API. We provide factual, verifiable details about our voice licensing, data handling, and infrastructure without marketing exaggeration. The information is intended for technical decision-makers and engineering leads who require precise specifications for integrating synthetic speech into SaaS notifications and automated interfaces.

You will find specific details on our voice sources, including the LibriTTS-R dataset and CC BY 4.0 licensing. We also clarify our data privacy policies, technical limitations such as the lack of an on-premise option, and our support structure. This ensures your team can accurately assess our API’s fit for high-volume, predictable audio generation workflows.

About SynthLogic and Our Text-to-Speech Infrastructure

Frequently asked questions

Who owns SynthLogic?

SynthLogic operates as an independent storefront for the SynthLogic text-to-speech API. Each site in the network functions as a separate entity. We are not affiliated with OpenAI, ElevenLabs, Google, Amazon, Microsoft, or any other brand mentioned in comparisons. Our structure is designed to provide direct access to the API infrastructure without intermediary branding.

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. They are not custom-cloned from specific individuals. Every generated file includes an inaudible watermark marking it as synthetic speech. We do not offer custom voice cloning or upload your own voice for replication.

Do you use customer data for training?

Input text is used solely to produce the audio response and is not used for training models. Generated audio is returned in the response and not stored. We retain request metadata, such as time, character count, and status, strictly for billing purposes. Your data remains isolated from our training datasets.

How do I contact support?

Support is available via email only. We do not offer a phone number, live chat, or direct voice support. You can reach our team through the contact page. Responses are handled asynchronously. Please include your API key and relevant request details to help us resolve your issue efficiently.

Is there a physical office?

We do not maintain a physical office for customer visits. As a cloud-based API provider, our operations are entirely digital. There is no on-premise install option or data-center choice for customers. All services, including billing and support, are managed remotely through our web interface and email channels.

Do you have a mobile app?

We do not offer a dedicated mobile app. You can access the API and manage your account via any web browser. Use the official OpenAI SDKs or plain HTTP requests to integrate the service into your mobile applications. We provide no mobile SDK, WordPress plugin, or browser extension.

Are the voices customizable?

The eight voice presets are fixed and not customizable. You cannot adjust emotion tags, phoneme dictionaries, or pronunciation lexicons. There is no SSML support. You can adjust the speed between 0.25 and 4.0 and select the output format. The voice characteristics remain constant across all generations for consistency.

Do you offer refunds?

Failed requests are free, with the charge returned in full if synthesis fails. Standard successful requests are non-refundable. We do not offer invoices with VAT or monthly subscriptions that require periodic billing adjustments. Top-ups via USDT on the TRON network (TRC-20) are credited automatically. Balance does not expire, but usage is final upon success.

Is the API open source?

The API itself is not open source. It is a proprietary cloud service. However, it uses the same request shape as the OpenAI speech endpoint, allowing you to use the official OpenAI SDKs by changing the base URL. We do not provide the source code for the synthesis engine.

Do you have a blog?

We do not maintain a public blog. Updates and technical details are provided through the documentation and API endpoints. We share product facts directly on the site pages. There are no editorial articles, case studies, or news posts. All information is structured around the API capabilities and pricing models.

Can I sue if the audio fails?

We do not offer an SLA or uptime guarantee. There is no enterprise contract defining liability limits. If a request fails, you receive a full refund of the character cost. We are not liable for indirect damages or business interruption caused by service latency or failures during high load periods.

Do you store my API key?

Your API key is stored securely on our servers to authenticate your requests. You can issue a new key at any time, which immediately invalidates the old one. We retain request metadata for billing but do not store the generated audio files. Your key is specific to your account on this site.

Is the site affiliated with OpenAI?

We are not affiliated with OpenAI, despite using a compatible endpoint structure. We are an independent storefront for the SynthLogic TTS API. We do not use OpenAI branding or infrastructure. Our voices, pricing, and rate limits are distinct. We are listed as an alternative provider, not a subsidiary or partner.

How do I report a bug?

Report bugs via email support. Include your API key, the request ID if available, and a description of the issue. We do not have a public issue tracker or bug bounty program. Our engineering team reviews reports asynchronously. Please provide the input text and voice used to help us reproduce the error.

Our Editorial Mission

Sound wave illustration for: Our Editorial Mission

SynthLogic operates as an independent storefront for the SynthLogic Text-to-Speech API. We provide direct access to the underlying engine without intermediary markup. Our interface displays real usage metrics and pricing data, allowing startups to integrate synthetic speech into SaaS notifications and automated interfaces with predictable costs.

We maintain a single active API key per account, ensuring straightforward billing. The endpoint structure mirrors standard industry conventions, allowing developers to use existing OpenAI SDKs by adjusting the base URL. This reduces integration friction for engineering teams building voice-enabled products.

Independent API Storefront

Independent API Storefront

Unlike resellers who bundle TTS into broader platforms, we focus exclusively on the API infrastructure. This specialization ensures that our documentation and pricing reflect the actual capabilities of the engine, including the 8 voice presets and 23 supported languages, without hidden service layers.

  • Direct access to the TTS endpoint
  • Transparent character counting and billing
  • Standard SDK compatibility

Fact-Based Reporting

Fact-Based Reporting

All performance metrics and feature lists are derived from live API behavior. We report exact character limits (4096 per request) and generation speeds based on real-time processing. This allows engineering teams to accurately estimate latency and throughput for their specific use cases.

  • Verified generation speeds
  • Accurate character limits
  • Real-time usage data

No Hype Policy

We avoid subjective claims like 'most realistic' or 'best'. Instead, we specify that voices are based on the LibriTTS-R dataset and LibriVox public-domain recordings. Our descriptions of voice characteristics—such as 'warm' for Aria or 'neutral' for Atlas—are based on consistent listener feedback, not marketing superlatives.

Voice Licensing and Sources

Sound wave illustration for: Voice Licensing and Sources

SynthLogic’s text-to-speech voices are derived from licensed recordings sourced from the LibriTTS-R dataset. This foundation ensures that our synthetic outputs maintain high acoustic quality while remaining rooted in verifiable, publicly accessible source material.

We provide clear attribution for the voice models used in our API. Understanding the origin of the audio data helps developers assess licensing compatibility for their own commercial applications.

LibriTTS-R Dataset

The LibriTTS-R dataset serves as the core training material for our eight voice presets. It consists of high-quality speech recordings extracted from the larger LibriSpeech corpus.

These recordings capture a wide range of intonations and speaking styles, which we distill into distinct voice profiles like Aria, Nova, and Atlas.

  • High-quality speech recordings for training
  • Distinct voice profiles derived from source data
  • No artificial voice design or modification

CC BY 4.0 License

The voices used in our API are based on recordings licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This allows us to legally integrate these recordings into our commercial TTS service.

Users of our API receive generated audio files that are synthetic derivatives. Every generated file carries an inaudible watermark that marks it as synthetic speech.

  • Based on CC BY 4.0 licensed recordings
  • Commercial use permitted under original license terms
  • Clear attribution for source material

LibriVox Public Domain

Many of the underlying recordings originate from LibriVox, a platform dedicated to public domain audiobooks. These recordings are typically in the public domain or released under permissive licenses, ensuring broad usability.

By leveraging these established public resources, we maintain a transparent and legally sound infrastructure for generating synthetic speech without relying on proprietary voice actor contracts.

  • Source recordings from LibriVox platform
  • Public domain and permissive license sources
  • Transparent legal foundation for voice models
Voice Licensing and Sources

Data Policy and Privacy

Sound wave illustration for: Data Policy and Privacy

SynthLogic processes your input text solely to generate the requested audio file. We do not retain your source text for training our models or for any secondary purpose. Generated audio is returned in the response and not stored.

We maintain request metadata—including timestamp, character count, and status—for billing and usage verification purposes only. This data supports accurate invoicing and rate-limit management without retaining the actual audio or text content. Your data usage is strictly limited to fulfilling the API request and accounting for charges.

Input Text Usage

Input text is used exclusively for synthesis and is not used for training. Generated audio is returned in the response and not stored.

Audio Storage Duration

Generated audio is returned in the response and not stored.

Metadata Retention

Request metadata (time, characters, status) is kept for billing.

Technical Infrastructure

Sound wave illustration for: Technical Infrastructure

SynthLogic operates as a cloud-based API provider, delivering text-to-speech services through standard HTTP requests. There is no on-premise installation option; all processing occurs in our infrastructure.

We use a standard endpoint compatible with existing SDKs, ensuring straightforward integration for SaaS applications without custom client requirements.

Cloud-Based Processing

All speech synthesis occurs in our cloud environment. Text is processed server-side, and audio files are returned directly in the response.

Requests are processed one after another. Under load, a request may wait in a queue. Typical speed is roughly real time for long texts, though short phrases may take a few seconds.

  • No local software to install
  • Server-side synthesis

Standard Endpoint

We use a standard POST endpoint at https://api.play76d.com/v1/audio/speech. The request shape matches the OpenAI speech endpoint, allowing you to use official OpenAI SDKs by simply changing the base URL.

This standardization simplifies integration for startups. You send text and receive an audio file in formats like mp3, wav, opus, flac, or pcm.

No On-Premise Option

We do not offer an on-premise installation or data-center choice. All data is processed in our cloud infrastructure.

This approach keeps costs predictable and removes the overhead of maintaining your own speech synthesis servers.

Technical Infrastructure

Support and Contact

Sound wave illustration for: Support and Contact

SynthLogic maintains a lean operational structure to keep infrastructure costs predictable for startups. Consequently, our support model is streamlined to ensure that technical inquiries are handled efficiently without the overhead of multi-channel communication teams.

We prioritize clear, documented resolution paths over real-time availability. This approach allows our engineering team to focus on API stability and voice engine performance, ensuring that your integration remains robust and reliable during high-volume usage.

Email Support Only

All support requests are managed exclusively through email. When you contact us, your issue enters a structured ticketing system where it is tracked from initial receipt to final resolution. This ensures that no inquiry is lost and that you receive a written record of the support interaction.

  • Standard response time within one business day
  • Written history of all support interactions for audit trails

No Live Chat

We do not offer live chat widgets on our site.

No Phone Number

There is no dedicated phone number for SynthLogic support.

Support and Contact

Limitations and Honesty

Sound wave illustration for: Limitations and Honesty

SynthLogic operates as an independent API storefront without enterprise-level service agreements. We do not provide an SLA guarantee for uptime, nor do we maintain certifications such as SOC 2 or HIPAA. Our infrastructure is designed for high-volume SaaS integration, prioritizing consistent character-level pricing over premium availability tiers.

No SLA Guarantee

We do not offer a formal Service Level Agreement (SLA) or uptime guarantee. Requests are processed sequentially, meaning under high load, a request may wait in a queue. This standard processing model ensures predictable character costs but does not promise instantaneous delivery during peak usage.

No Certifications

Our service does not hold specific compliance certifications like SOC 2, HIPAA, or GDPR certificates. Input text is used solely for audio synthesis and not for training models, but we do not provide enterprise contracts or on-premise installation options for regulated data environments.

Standard Queue Processing

Audio generation is handled via cloud-based processing with standard queue mechanics. While long texts generate at roughly real-time speed, short phrases may take a few seconds to complete. We do not offer streaming or WebSocket connections for real-time partial audio delivery.

Listen to the 8 voices

Each sample below was generated by this API from the same sentence. Transform your startup's user experience with seamless notifications and intuitive interfaces that keep teams connected and informed.

Avatar of the voice aria
ariafemale
Avatar of the voice nova
novafemale
Avatar of the voice luna
lunafemale
Avatar of the voice sage
sagefemale
Avatar of the voice felix
felixmale
Avatar of the voice milo
milomale
Avatar of the voice atlas
atlasmale
Avatar of the voice orion
orionmale

23 languages

Any voice speaks any of these languages.

Sources and further reading

  1. Wikipedia
  2. Node.js documentation
  3. Python documentation
  4. OpenAI API reference: create speech
  5. curl documentation
  6. LibriTTS-R paper (arXiv 2305.18802)
  7. Creative Commons Attribution 4.0
  8. openai-node on GitHub
  9. W3C WAI: making audio and video accessible
  10. IETF RFC 7845: Ogg Opus
  11. Tether (USDT)
  12. ISO 639 language codes

More on this site