Frequently asked questions
What does Reddit mean by voice cloning?
Voice cloning typically involves training a model on specific audio samples to replicate a unique speaker’s timbre and style. It allows for generating speech that sounds identical to a source person, often used for personalized content or recreating a specific voice without the original speaker’s continuous presence.
Does this API offer custom voice cloning?
No. We do not offer custom voice cloning, voice design, or the ability to upload your own voice. The API provides eight fixed presets based on licensed recordings. Users select from these existing voices rather than creating a unique clone for their specific needs.
How do preset voices differ from cloned voices?
Preset voices are fixed models derived from licensed recordings like LibriTTS-R. They offer consistent tone and style without requiring training data. Cloned voices are generated from specific user-provided audio samples to mimic a particular individual. Our eight presets provide reliable, standardized options without the variability of cloning.
Can I use these voices for commercial dubbing?
Yes. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobooks. Generated audio files carry an inaudible synthetic watermark. There are no restrictions on commercial use, making these voices suitable for professional dubbing workflows and global content localization.
Are the voices based on real people's recordings?
Yes. Every voice is based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain readings from LibriVox. These are real human voices recorded for audiobooks. Each generated file includes an inaudible watermark to identify it as synthetic speech derived from these source recordings.
What is the character limit per request?
Each request accepts up to 4096 characters of input text. Spaces and punctuation count toward this limit. For longer texts, you must split the content into parts of 4096 characters or less and join the resulting audio files manually or via your integration logic.
How many languages are supported?
The API supports 23 languages, including English, Spanish, German, French, Japanese, Korean, Chinese, Arabic, Hindi, and Russian. Any voice can speak any of these languages. If the language parameter is omitted, the system guesses from the script, defaulting to English for Latin text.
Is there a free trial available?
Yes. Signing in with Google provides $0.50 of credit, valid for 7 days, covering approximately 50,000 characters. Accounts created with email and password start without trial credit. You can top up via USDT (TRC-20) to continue using the API after the trial expires.
What payment methods are accepted?
We accept USDT on the TRON network (TRC-20) only. Minimum top-up is $5. Credits are added automatically within minutes. We do not accept credit cards, PayPal, or subscriptions. There are no invoices with VAT. Balance does not expire, and failed requests are refunded.
Does the API support streaming audio?
No. The API returns the complete audio file upon request completion. There is no streaming, WebSocket, or real-time partial audio support. You receive the final audio file via the POST response. This is suitable for batch processing but not for live, real-time voice interaction.
Is SSML supported for fine control?
No. The API does not support SSML, phoneme dictionaries, or custom pronunciation lexicons. Control is limited to voice selection, speed (0.25–4.0), and language code. For precise timing or pronunciation, you must manage these aspects within your text input or post-processing audio tools.
How long are requests queued during high load?
Requests are processed sequentially. Under load, a request may wait in a queue. Typical speed is roughly real-time (25 characters per second). Short phrases take a few seconds. Long texts are split and joined. There is no SLA or uptime guarantee to specify exact queue times.
What happens if a synthesis request fails?
If synthesis fails, the charge is returned in full. You are not billed for failed requests. This ensures you only pay for successful audio generation. Failed requests do not consume your character limit or balance, providing a reliable billing model for your localization projects.
Do I need a specific SDK to use the API?
No. You can use the official OpenAI SDKs by changing the base URL to https://api.vs55.info/v1. The request shape matches the OpenAI speech endpoint. You can also use plain HTTP or any standard HTTP client. We do not provide a proprietary SDK or mobile SDK.
First request
curl
curl https://api.vs55.info/v1/audio/speech \
-H "Authorization: Bearer $TTS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"tts-1","voice":"aria","input":"Hello from the API.","response_format":"mp3"}' \
--output hello.mp3Understanding Synthetic Speech in Localization
Synthetic speech has evolved from robotic readouts to nuanced narration suitable for global localization. Discussions often conflate any AI voice with "cloning," but true voice cloning requires training on unique samples. Our API uses eight fixed presets derived from licensed recordings, ensuring consistent quality without the variability of custom models.
For dubbing agencies, this distinction matters. Presets offer predictable character profiles and instant availability, whereas custom cloning demands specific source material and longer turnaround times. We evaluate these models to help agencies choose the right tool for their localization pipeline.
What Reddit Means by Voice Cloning
Online communities frequently use "voice cloning" as a catch-all term for any AI-generated speech. However, true cloning involves training a model on a specific individual's voice data. Our API does not offer this; instead, it provides eight distinct, pre-trained voice presets that simulate different tones and styles.
- No custom voice cloning available
- Fixed library of 8 presets
- Consistent output per voice
Preset Voices vs. Custom Cloning
Preset voices like Aria or Atlas are based on licensed recordings from LibriTTS-R and LibriVox. They are ready for immediate use via API. Custom cloning, by contrast, requires uploading unique audio samples and waiting for training. For most localization projects, presets provide sufficient variety without the overhead of custom model development.
Why Agencies Prefer Presets
Agencies favor presets for their speed and cost-efficiency. With a single API key, you can generate thousands of hours of content across 23 languages without negotiating individual voice licenses. The predictable billing structure ($0.01 per 1,000 characters) and lack of training delays make presets ideal for high-volume dubbing workflows.
API Capabilities for Dubbing Workflows
The core of our TTS API is a single POST endpoint that converts input text into audio files. By aligning with the OpenAI request shape, agencies can integrate the service into existing localization pipelines using standard HTTP clients or the official OpenAI SDK, simply by adjusting the base_url to https://api.vs55.info/v1. This compatibility ensures a straightforward implementation for developers.
Text-to-Speech Endpoint Overview
The endpoint accepts JSON bodies containing the model (tts-1 or tts-1-hd), input text, voice preset, response_format, and optional speed or language parameters. Authentication is handled via a Bearer token, allowing immediate integration into automated dubbing workflows. The API processes requests sequentially, meaning longer texts should be split into chunks of 4096 characters or fewer to ensure consistent delivery.
- Endpoint: POST https://api.vs55.info/v1/audio/speech
- Auth: Authorization: Bearer tts-...
- Rate limit: 60 requests/minute per key
- Input limit: 4096 characters per request
Supported Languages for Global Content
The API supports 23 languages, covering major markets for localization. Voices are versatile; any preset can speak any of the supported languages, allowing agencies to maintain consistent vocal characteristics across different language versions of a project. If the language field is omitted, the system guesses the language from the script, defaulting to English for Latin text.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Audio Formats for Post-Production
For post-production flexibility, the API offers five audio formats: mp3, wav, opus, flac, and pcm. These formats cater to different stages of the dubbing workflow, from quick previews to lossless archival. Generated audio is returned directly in the response and not stored, ensuring that your processing pipeline remains lean and focused on immediate editing needs.
Voice Presets and Character Profiles
Dubbing agencies rely on consistent, licensed voices for localized content. Our API provides eight distinct presets derived from LibriTTS-R and LibriVox recordings, ensuring clear, professional audio without the legal complexities of custom voice cloning.
Female Voices: Aria, Nova, Luna, Sage
Aria delivers a warm, clear narration. Nova offers a bright, upbeat tone for promotional content. Luna provides a soft, storytelling quality for narrative-driven projects. Sage ensures calm, instructional clarity for educational materials.
Male Voices: Felix, Milo, Atlas, Orion
Felix brings energetic delivery suitable for ads. Milo offers a friendly, conversational style for conversational content. Atlas maintains a neutral, news-ready tone. Orion provides a deep voice.
Voice Suitability for Narration
Each voice supports all 23 languages, allowing consistent character selection across global releases. Agencies can select presets that match specific tonal requirements without needing custom training or additional licensing steps.
Audio files are returned in standard formats (mp3, wav, opus, flac, pcm) for seamless integration into post-production workflows. The synthetic watermark ensures transparency without affecting audio quality.
- Aria: Warm, clear narrator
- Nova: Bright, upbeat
- Luna: Soft, storytelling
- Sage: Calm, instructional
- Felix: Energetic, promo
- Milo: Friendly, conversational
Pricing Models and Cost Efficiency
Dubbing agencies often calculate costs based on character volume rather than audio duration, making per-character billing a transparent choice for budgeting localization projects. This model ensures you pay only for the text processed, with spaces and punctuation counting toward the total.
The structure is straightforward: $0.01 per 1,000 characters. Failed requests incur no charge, protecting your budget during testing phases. This predictability helps agencies estimate total dubbing costs before committing to large-scale voiceover production.
Per-Character Billing Structure
Billing is strictly per character, including spaces and punctuation. This unit allows precise cost estimation for scripts of varying lengths. Failed synthesis attempts are refunded in full, ensuring you only pay for successful audio generation.
This model aligns well with localization workflows where text volume is the primary variable. Usage scales without monthly fees or seat counts.
- Spaces and punctuation count as characters.
- Failed requests are free.
- No monthly plans or minimum spend.
Free Trial and Top-Up Options
New accounts created with Google sign-in receive a $0.50 trial credit, valid for 7 days. This provides approximately 50,000 characters for initial testing. Email-based accounts start with zero balance.
Top-ups are processed via USDT on the TRON network (TRC-20). The minimum deposit is $5, and funds are credited automatically within minutes. There are no credit cards, PayPal, or subscriptions to manage.
- $0.50 trial for Google sign-ins.
- Minimum top-up: $5 USDT.
- Auto-credit via TRC-20 within minutes.
Cost Per Million Characters
The rate is $10 per 1 million characters, or $0.01 per 1,000 characters. This flat rate applies regardless of the voice model (tts-1 or tts-1-hd) selected.
Balance does not expire, allowing agencies to accumulate credits for future projects. The simplicity of the pricing model reduces administrative overhead for finance teams managing multiple localization campaigns.
- $10 per 1 million characters.
- $0.01 per 1,000 characters.
- Balance never expires.
Technical Limits and Performance
The 60 requests-per-minute limit per key dictates throughput. New accounts start at 10 requests/minute until the first top-up, requiring careful batching of long-form content to avoid throttling during peak production windows.
Request Rate Limits Explained
Each API key supports a maximum of 60 requests per minute. This hard limit applies regardless of voice or format selection. Agencies processing multiple language variants simultaneously must distribute load across keys or stagger requests to maintain steady audio generation without hitting the ceiling.
Text Length Constraints (4096 Chars)
The input field accepts up to 4,096 characters per request. Scripts exceeding this length must be split into sequential parts. The client application must join the resulting audio files to ensure seamless playback, as the API returns discrete audio chunks rather than a single continuous stream.
Generation Speed and Queue Times
Synthesis typically matches real-time speed, processing roughly 25 characters of text per second of audio generated. Short phrases may take a few seconds to complete. Under heavy load, requests are processed sequentially, meaning a request may wait in a queue before synthesis begins, extending total turnaround time for large batches.

Data Privacy and Training Policy
Data handling in our TTS API is designed for agency workflows where input privacy matters. Text is processed solely to generate the requested audio file and is not used for model training or improving voice synthesis algorithms. This ensures that your localized scripts remain proprietary during the production phase.
Input Text Usage for Synthesis
Input text is transmitted to the engine, converted to audio, and then discarded. We do not retain the actual text content after generation is complete, ensuring that your dialogue scripts are not ingested into training datasets. This policy applies to all characters processed, regardless of volume.
Audio Storage and Retrieval
Generated audio files are returned directly in the API response and are not stored on our servers by default. Once you receive the file, we hold no copy of the audio. This reduces data residency risks and keeps your final deliverables entirely under your control.
Billing Metadata Retention
We retain request metadata strictly for billing purposes. This includes the timestamp, character count, and request status (success or failure). This data is used only to calculate costs and process refunds for failed requests, and is not used for analytics or training.
Licensing and Synthetic Watermarks
Synthetic voices are derived from licensed recordings, ensuring a clear legal foundation for commercial localization projects. Understanding the source material helps agencies navigate copyright considerations when deploying generated audio in global dubbing pipelines.
LibriTTS-R and CC BY 4.0 Sources
Voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. This source material is available under the CC BY 4.0 license, providing a transparent and verifiable origin for the synthetic voice models used in the API.
Inaudible Audio Watermarks
Every generated audio file carries an inaudible watermark that explicitly marks the content as synthetic speech. This embedded identifier allows for easy verification of the audio's origin without affecting the listening experience or requiring additional metadata files.
Legal Implications for Distribution
The combination of CC BY 4.0 sourced models and synthetic watermarks clarifies usage rights for distributors. Agencies can confidently deploy these voices in commercial content, knowing the synthetic nature is both legally grounded and technically detectable, reducing ambiguity in rights management.
Comparison with Other TTS Solutions
Dubbing agencies often require seamless integration with existing engineering pipelines. Our API uses the exact request shape of the OpenAI speech endpoint, allowing teams to switch providers by simply updating the base URL in their OpenAI SDK configurations without rewriting code.
OpenAI SDK Compatibility
The API endpoint matches the OpenAI structure precisely. This means you can use official OpenAI SDKs by changing the base_url to https://api.vs55.info/v1 and using the site's key. No custom client libraries or complex middleware are required for integration.
No Custom Voice Cloning Feature
Unlike specialized cloning services, we do not offer custom voice cloning or uploading of proprietary voice samples. We provide eight high-quality presets derived from licensed LibriTTS-R recordings. This ensures consistent, legally clear audio without the variability of neural cloning.
- No uploading of custom voice samples
- No per-voice training required
- Consistent preset quality across all languages
Simple HTTP vs. Complex SDKs
We rely on standard HTTP requests rather than proprietary SDKs. This reduces dependency lock-in and simplifies debugging. Agencies can use plain HTTP or any standard library, avoiding the overhead of maintaining multiple language-specific SDKs.
Editorial Assessment for Agencies
Preset voices offer a reliable, cost-effective baseline for dubbing, ensuring consistent quality across high-volume projects without the overhead of custom voice design. Their broad language support and predictable pricing make them ideal for standard localization tasks.
When to Use Preset Voices
Use preset voices for standard narration, e-learning modules, and general content where specific brand alignment isn't critical. The eight available voices cover a wide range of tones, from warm narration to energetic promo, making them suitable for most bulk localization needs.
- High-volume standard content
- Budget-conscious projects
- Rapid prototyping
When to Hire Human Actors
Hire human actors when emotional nuance, specific character branding, or unique timbre is essential. Preset voices lack the subtle inflections and personalized delivery that human performers provide, making them less suitable for premium, emotionally driven campaigns.
Summary of API Strengths
The API offers straightforward HTTP integration and transparent per-character billing. Its compatibility with OpenAI SDKs and broad language support make it a practical tool for agencies seeking reliable, scalable synthetic speech solutions.
- Simple HTTP integration
- Transparent billing
- Broad language support
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Split logic should respect word boundaries to avoid cutting mid-sentence.
23 languages
Any voice speaks any of these languages.
