Frequently asked questions
What is TTS AI voice?
LexiRead is a cloud text-to-speech API that converts text into audio files via a simple POST request. Optimized for e-commerce startups, it generates product descriptions and notifications efficiently without enterprise bloat. By integrating directly into your workflow, you can quickly produce audio content using our straightforward API endpoint.
How does text to speech convert text to audio?
You send text via a JSON POST request to the API endpoint. The system processes the input and returns an audio file in formats like MP3, WAV, or FLAC. For texts over 4096 characters, split them into parts and join the resulting audio files. Requests are processed sequentially, so load may cause queuing.
Are the voices real humans or AI generated?
The voices are AI-generated based on licensed recordings from LibriTTS-R and LibriVox public-domain audiobook readings. Every generated file carries an inaudible watermark marking it as synthetic speech. They are not real-time recordings of specific individuals but synthesized audio created from these licensed source materials.
Can I clone a specific voice?
No, custom voice cloning is not offered. You cannot upload your own voice or design custom voices. You must choose from the eight available presets (e.g., Aria, Felix, Atlas). There is no voice cloning feature for customers, so you cannot replicate specific human voices beyond the provided options.
Does the AI understand context?
The API does not support SSML, emotion tags, or phoneme dictionaries. It converts text directly to audio without advanced contextual understanding or custom pronunciation lexicons. You control tone via voice selection and speed settings. For complex formatting, you must manage it within your text input before sending the request.
Is the audio quality comparable to human voice?
Our voices are based on licensed LibriTTS-R recordings (CC BY 4.0) from LibriVox audiobooks. They offer clear, natural-sounding speech suitable for e-commerce product descriptions. While not identical to real humans, they avoid the robotic feel of older TTS systems and include an inaudible watermark to mark them as synthetic speech.
How many languages are supported?
We support 23 languages, including English, Spanish, German, French, Chinese, Japanese, and Arabic. Any voice can speak any of these languages. If the language is not set, it guesses from the script, defaulting to English for Latin text. Vietnamese, Indonesian, and Thai are not supported.
Can I use the audio commercially?
Yes. The voice models are based on public-domain LibriVox recordings, allowing commercial use in product descriptions and notifications. Input text is not used for training, and generated audio is returned in the response without long-term storage by the API. This makes the service suitable for e-commerce notifications and product content.
Is there a delay in generation?
Typical speed is roughly real-time, about 25 characters per second. Short phrases take a few seconds. Requests are processed one after another, so under load, a request may wait in a queue. Longer texts require splitting and joining, which adds processing time compared to single requests.
Do I need a special player for the audio?
No, the API returns standard audio formats like MP3, WAV, Opus, FLAC, or PCM. These are universally supported by web browsers and mobile devices. You can embed the audio directly in your e-commerce store or send it via notifications without needing a proprietary player or SDK.
Can I control the speed of the speech?
Yes, you can adjust the speed between 0.25 and 4.0 using the speed parameter in your request. This allows you to speed up product details or slow down instructions. The default speed is 1.0. This control helps tailor the audio to different e-commerce use cases effectively.
What is the difference between tts-1 and tts-1-hd?
Both models use the same engine and price. The only difference is the model name you send in the request. You can use either "tts-1" or "tts-1-hd" without any change in cost or audio quality. They are functionally identical for your e-commerce integration needs.
Is the watermark audible to users?
No, the watermark is inaudible. It marks the file as synthetic speech but does not affect the listening experience. Your customers will hear clear, natural-sounding voiceovers for your products without noticing any hidden audio markers or artifacts added by the synthesis process.
Can I integrate this with Shopify or WooCommerce?
Yes, you can integrate via the standard HTTP API or OpenAI SDKs. Since it uses the same request shape as OpenAI, you can change the base_url to our API endpoint. This allows easy integration into custom e-commerce platforms, allowing you to generate audio for product descriptions programmatically.
Definition of TTS AI Voice
Text-to-speech technology converts written content into audible speech via cloud APIs. You send a text payload and receive an audio file, making it ideal for automating product descriptions and notifications in online stores without hiring voice talent.
Text to Speech Basics
The system accepts text input and returns synthesized audio files. Unlike human recording, this process scales effortlessly, allowing you to generate thousands of unique audio files from a single script without repetitive studio sessions.
Neural Voice Synthesis
Our voices are built from licensed recordings of LibriTTS-R, sourced from public-domain audiobooks. This foundation ensures natural, high-quality speech without the need for complex neural network training on your specific data.
Synthetic Watermarks
Every generated audio file carries an inaudible watermark. This marks the content as synthetic speech, providing transparency for your customers while keeping the listening experience clean and professional.
How the Engine Works
The engine accepts text input (up to 4096 characters) and processes it into audio files. You send a JSON payload to the API endpoint, specifying the model and voice. The system returns a standard audio file, with no streaming or WebSocket connections required for basic operation.
Input Processing
Text is validated for length and character count. Spaces and punctuation are counted toward the total. If the input exceeds limits, split the text into chunks. The API guesses the language from the script, but you should set the language parameter for non-English Latin text to ensure accuracy.
Audio Generation Steps
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 to generate. Failed requests incur no charge.
Real-time vs Batch
Unlike services offering live streaming, this API returns complete audio files after processing. There is no WebSocket or partial audio stream. You receive the final file via the response, making it suitable for asynchronous workflows like email notifications or stored product pages.

Voice Selection Criteria
Selecting the right voice ensures your audio content aligns with your brand's tone. The API provides eight distinct presets, allowing you to match specific use cases, from energetic promotions to calm instructions.
Tone and Style
Choose from voices like Felix for energetic promos or Sage for calm instructions. Aria offers a warm, clear narration suitable for detailed product descriptions, while Nova provides a bright, upbeat tone for engaging audio banners.
Gender and Age Range
Available voices are categorized by gender: four female presets (Aria, Nova, Luna, Sage) and four male presets (Felix, Milo, Atlas, Orion). Specific age ranges are not defined; selection is based on vocal character rather than demographic markers.
Narrative vs Conversational
Use Luna for soft, storytelling narratives or Milo for friendly, conversational order confirmations. Atlas provides a neutral, news-style delivery for straightforward updates, ensuring clarity and professionalism in customer notifications.
Use Cases in E-commerce
TTS AI converts text into speech, enabling automated audio for product details, alerts, and banners. By splitting long inputs into 4096-character segments, the API joins audio seamlessly, ensuring complete delivery despite sequential processing queues.
Product Descriptions
Convert static product text into engaging audio. With a generation rate of ~25 characters per second, long descriptions play at real-time speed. The system splits text into 4096-character chunks, joining the resulting audio files to maintain continuity.
Order Confirmation Alerts
Generate audio notifications for order updates. Requests are processed sequentially, so batch processing ensures reliable delivery without overwhelming the queue. Note that delivery is synchronous but not instant; generation time depends on text length and current load.
Audio Banners
Create short audio snippets for website banners. These use the same API endpoint, allowing consistent voice selection across product pages and marketing elements. Ideal for brief, high-impact messages where text might be overlooked.
Technical Requirements
Integration leverages the standard OpenAI SDK by pointing the base_url to https://api.getking11.com/v1. This allows you to use familiar libraries without learning a proprietary client. Authentication relies on a single Bearer key, simplifying header management across your backend services.
API Integration
Send a POST request to /v1/audio/speech with JSON parameters. The API accepts text up to 4096 characters per call. For longer product descriptions, split the input and concatenate the resulting audio files. Requests process sequentially, so high-volume batches may queue under load.
- Use the official OpenAI SDK by overriding
base_url. - Auth via
Authorization: Bearer tts-...header. - Sequential processing may cause slight delays during peak loads.
Bandwidth Considerations
Data transfer volume depends directly on the selected response_format and audio duration. Longer texts generate larger files, particularly in WAV or FLAC. Plan your egress limits based on the total character count multiplied by the format's bitrate, rather than assuming minimal usage.
- MP3 and Opus offer smaller file sizes for bandwidth efficiency.
- WAV and FLAC provide lossless quality but increase transfer costs.
- PCM returns raw data, requiring significant bandwidth for high-quality output.
Storage Costs
The API returns audio in the response but does not store it permanently. You are responsible for storing generated files if needed for caching or audit trails. Storage costs will depend on your cloud provider's pricing for the volume of audio assets retained.
Comparison with Human Voiceovers
Manual voiceovers require studio time, talent fees, and editing for each script change, creating a linear cost curve. Our API charges only $0.01 per 1,000 characters, with no minimum spend or monthly seats. Failed requests are free, ensuring you pay strictly for successful audio generation.
Cost Efficiency
Human voiceovers incur fixed costs for talent and studio time, regardless of volume. The API scales linearly: $10 per million characters. You avoid overhead for minor text tweaks, as regeneration is instant and cheap. This model suits startups with tight margins who need frequent updates to product descriptions or notifications without renegotiating contracts.
- No talent fees or studio rental costs.
- Pay only for successful synthesis ($0.01/1k chars).
- Zero cost for failed requests.
Scalability
While manual voiceovers are limited by human availability, the API handles volume through sequential processing. Requests are processed one after another, so under load, a request may wait in a queue. Scale is immediate in terms of access, but throughput is bounded by a 60 requests-per-minute limit per key. This is sufficient for most e-commerce catalogs but requires queuing logic for massive bursts.
- Process thousands of SKUs without hiring more talent.
- Rate limit: 60 requests/minute per key.
- Sequential processing means queuing under high load.
Consistency
Human voice quality varies by session, mood, or accent drift. Our eight presets (e.g., Aria, Atlas) deliver identical tone every time. Each generated file also carries an inaudible synthetic watermark, ensuring brand consistency and authenticity. This reliability is critical for standardized order confirmations and global product descriptions where brand voice must remain uniform across all touchpoints.
Limitations of Current Tech
We analyze TTS APIs for e-commerce startups, focusing on cost efficiency and integration speed for product descriptions and notifications. Our reviews help you choose the right voice engine without enterprise bloat.
No Emotion Tags
Unlike human narrators, the engine cannot dynamically adjust tone based on context. You cannot tag a sentence with 'excitement' or 'sadness' to alter the delivery. The voice remains consistent, which is ideal for uniform product descriptions but less suitable for dramatic storytelling.
No Custom Cloning
The service uses eight fixed presets rather than allowing brand-specific voice cloning. You cannot upload a sample of your CEO’s voice to create a custom brand asset. Each generated file carries an inaudible watermark identifying it as synthetic speech, ensuring transparency.
No SSML Support
The API does not support Speech Synthesis Markup Language (SSML). You cannot control pause durations, pronunciation, or emphasis via XML tags. Control is limited to the speed parameter (0.25–4.0). This simplifies integration but removes granular audio pacing control.

Language Support Details
LexiRead supports 23 languages, allowing any of the 8 voice presets to speak in any supported language. This flexibility lets e-commerce startups localize product descriptions and notifications without hiring separate voice talent for each market.
Supported Languages List
You can generate audio in Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. The API guesses the language from the script if not specified, though setting the language parameter is recommended for non-English Latin text to ensure accuracy.
Unsupported Languages
Vietnamese, Indonesian, Thai, Ukrainian, Romanian, Czech, and Bengali are not currently supported. If your store targets these regions, you will need a different solution or manual recording for those specific audio assets.
Cross-Lingual Capabilities
Any voice can speak any of the 23 supported languages. This means you can maintain a consistent brand voice across different markets by using the same voice preset, such as Aria or Atlas, for both English and German product descriptions.
Audio Formats Available
The API supports five output formats: mp3, wav, opus, flac, and pcm. You select the response_format parameter in your request to match your e-commerce player’s capabilities. MP3 offers the best compression for general web delivery, while WAV provides uncompressed quality for local processing.
MP3 vs WAV
MP3 is ideal for reducing bandwidth costs on product description pages, where file size matters. WAV files are larger but ensure maximum fidelity for high-end audio players or local storage without re-encoding artifacts.
OPUS and FLAC Options
OPUS provides efficient compression suitable for mobile apps and low-bandwidth environments. FLAC offers lossless compression, preserving audio quality while reducing storage needs compared to raw WAV files.
PCM Raw Data
PCM returns raw uncompressed audio data. This format is useful for immediate processing in custom audio pipelines or when you need to apply additional effects without decoding overhead.
Data Privacy Implications
Our API handles your e-commerce text data with strict privacy boundaries. Input text is processed solely to generate the audio file and is not retained for model training or general use. This ensures that unique product descriptions and customer notification content remain confined to the specific generation session.
Generated audio is returned directly in the response body and is not stored on our servers. Only request metadata—such as timestamp, character count, and status—is kept briefly for accurate billing. This minimal data retention reduces liability and keeps your customer communication data secure and transient.
Input Text Usage
Text sent to the API is used exclusively for synthesizing the requested audio clip. Once the audio is generated and delivered, the input text itself is not archived for long-term storage or analysis, ensuring your product copy remains private.
Audio Storage Policy
The synthesized audio file is delivered instantly in the HTTP response and is not saved on our infrastructure. This eliminates storage costs for you and ensures that no audio files persist on our servers after delivery.
Training Data Exclusion
Your input text and generated audio are not used to train our voice models. This means your unique voiceovers and product descriptions remain distinct from the data used to improve the underlying engine, preserving your brand's originality.
Future of TTS in Retail
TTS AI voice continues to evolve as a practical tool for e-commerce startups seeking scalable audio content. While true streaming remains unavailable, generation speed approaches real-time, enabling near-instant audio creation for product descriptions and notifications. This efficiency supports rapid content updates without enterprise-level infrastructure.
The technology’s consistency and low cost make it ideal for high-volume retail environments. Startups can leverage this to maintain brand tone across thousands of SKUs, ensuring uniform customer experience without the variability of human voice actors.
Real-time Personalization
Although the API does not support WebSocket streaming, the generation speed is roughly real-time for long texts, allowing quick delivery of audio files. This enables dynamic insertion of personalized product details into order confirmations or browsing sessions without significant latency.
For short phrases, generation takes a few seconds, which is sufficient for pre-recorded audio banners or notification alerts. The lack of partial audio streaming means the full file is generated before delivery, but the speed remains competitive for batch and semi-real-time use cases.
- Roughly real-time generation speed for long texts
- No WebSocket streaming support
- Suitable for pre-recorded notifications and banners
Voice Branding
With eight distinct voice presets ranging from warm narrators to energetic promos, e-commerce brands can select a consistent voice for all product descriptions. This ensures that customers hear the same tone across different products, reinforcing brand identity without the need for custom voice cloning.
Voices are based on licensed recordings, providing a legal and ethical foundation for commercial use. The inaudible watermark in generated files helps identify synthetic speech, adding transparency to audio content.
- Eight distinct voice presets available
- Consistent tone across all SKUs
- Licensed recordings with synthetic watermarks
Integration Trends
The API’s compatibility with OpenAI SDKs simplifies integration for developers familiar with standard HTTP clients. By changing the base URL, existing e-commerce platforms can quickly add TTS capabilities to product pages or customer notifications without building custom infrastructure.
Support for multiple audio formats, including MP3, WAV, OPUS, and FLAC, ensures compatibility with various web players and mobile apps. This flexibility allows retailers to optimize file sizes for bandwidth efficiency while maintaining audio quality.
- OpenAI SDK compatibility
- Multiple audio format support
- Simplified integration via HTTP clients

Getting Started Checklist
Start by creating an account on getking11.com/get-api-key/ using Google or email. You receive a free $0.50 trial credit valid for 7 days, sufficient for approximately 50,000 characters. Your unique API key appears immediately for use.
Generate your first audio file by sending a POST request to https://api.getking11.com/v1/audio/speech. Use the provided curl example to test voice selection and format. Failed requests are free, allowing safe testing before full integration.
Sign Up Process
Visit the /get-api-key/ page and authenticate via Google or email/password. The dashboard instantly displays your active API key, current balance, and a ready-to-use curl example. No credit card is required for the initial trial credit of $0.50.
- Google or email authentication
- Instant API key generation
- $0.50 trial credit included
- No subscription required
First Request
Send a JSON payload to the /v1/audio/speech endpoint. Ensure the input text is under 4,096 characters. Specify a voice like 'aria' or 'nova' and choose a format such as mp3 or wav. The API returns the audio file directly in the response.
- POST to /v1/audio/speech
- Max 4,096 characters per request
- Supported formats: mp3, wav, opus, flac, pcm
Integration Testing
Verify the audio output and check for the inaudible synthetic watermark. Test speed variations between 0.25 and 4.0. Monitor your dashboard for character consumption. Remember that longer texts should be split into chunks and joined programmatically for seamless playback.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Build your online stores to scale quickly and attract new customers.
23 languages
Any voice speaks any of these languages.
