Frequently asked questions
What is a text-to-speech API?
A text-to-speech API is a cloud service that converts written text into audio files. You send text via a POST request to our endpoint and receive an audio file in return. It uses standard JSON bodies, allowing you to use official OpenAI SDKs by simply changing the base URL to our API domain.
How does TTS help language learners?
Our API provides high-quality synthetic voices in 23 languages, ideal for powering interactive exercises, flashcards, and reading assistants. Indie developers can use these voices to create immersive language learning tools without managing complex infrastructure, ensuring learners hear accurate pronunciation in their target language.
Which languages are supported?
We support 23 languages: Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. Any voice can speak any of these languages, and the system guesses the language from the script if not specified.
Can I use the voices for commercial apps?
Yes, you can use the generated audio in your commercial applications. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings. Each generated file carries an inaudible watermark marking it as synthetic speech, but there are no restrictions on commercial usage.
How fast is the API response time?
Typical speed is roughly real time, generating about 25 characters of text per second of audio for long texts. Short phrases may take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue before synthesis begins.
Do I need to split long texts manually?
You can send up to 4096 characters per request. For longer texts, you should split the input into parts of up to 4096 characters, send each part, and then join the resulting audio files. This ensures optimal processing and avoids exceeding the character limit per request.
Is there a free trial available?
Yes, signing in with Google gives you a free trial with $0.50 of credit, valid for 7 days, which covers about 50,000 characters. Accounts created with email and password do not receive this trial credit. The trial allows you to test the API before purchasing credits.
What audio formats do you support?
You can choose from mp3, wav, opus, flac, and pcm formats by setting the response_format field in your request. These formats cater to different use cases, from web playback to high-quality archival storage, ensuring compatibility with your specific language learning application requirements.
How do I get my API key?
Sign up on our /get-api-key/ page using Google or an email and password. Once registered, your account dashboard displays your API key, balance, usage statistics, and a ready-to-use curl example. The key is used for authentication via the Authorization header in your requests.
Can I change my API key later?
Yes, you can issue a new API key at any time from your account dashboard. When a new key is issued, the old one stops working immediately. This allows you to rotate keys for security purposes without interrupting your service if you use the new key.
What happens if a request fails?
If synthesis fails, the charge is returned in full. Failed requests are free, so you are not billed for unsuccessful attempts. This ensures you only pay for successfully generated audio files, protecting your budget when testing different voices or parameters.
Are the voices real humans or synthetic?
The voices are synthetic, generated from licensed recordings of human speakers. Each file carries an inaudible watermark marking it as synthetic speech. The voices are based on LibriTTS-R (CC BY 4.0) and LibriVox public-domain audiobook readings, providing high-quality natural-sounding speech without being direct human recordings.
Do you charge for failed requests?
No, failed requests are free. If synthesis fails for any reason, the charge is returned in full. You are only charged for successful audio generation, ensuring you pay strictly for the usable audio files delivered to your application.
How many characters can I send per request?
You can send up to 4096 characters per request. The input field accepts text up to this limit, including spaces and punctuation. For texts longer than 4096 characters, you must split them into multiple requests and join the resulting audio files.
First request
curl
curl https://api.kt10.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.mp3Why Language Learning Apps Need Reliable TTS
Language learning apps demand precise audio to help learners master pronunciation and rhythm. Our cloud API delivers high-quality synthetic voices in 23 languages, ensuring every word sounds natural and accurate for educational contexts.
Instant Pronunciation Models
Deliver accurate pronunciation for any text input, though generation takes time (roughly real-time speed). Short phrases take a few seconds, while longer texts generate at approximately 25 characters per second.
Consistent Voice Identity
Maintain a familiar auditory experience across your app. Users hear the same voice for flashcards, stories, and lessons, reinforcing recognition without the distraction of shifting tones.
Scalable Character Limits
Handle diverse content lengths easily. Process texts up to 4,096 characters per request, splitting longer lessons into seamless audio segments while keeping your infrastructure simple.
8 Voices for Diverse Learning Styles
Your app’s personality hinges on voice selection. We offer eight distinct presets based on licensed recordings from LibriTTS-R and LibriVox, ensuring your learners hear consistent, synthetic audio across all exercises.
Each voice carries a specific tonal character, allowing you to match the audio to the pedagogical goal—whether it’s quiet reading or active recall.
Warm Narrators for Stories
Use Aria or Luna for immersive reading passages. Aria provides a warm, clear narration style ideal for standard texts, while Luna offers a soft, storytelling tone that keeps learners engaged during longer sessions.
These voices maintain a gentle pace, making them perfect for beginner-level stories where clarity outweighs speed.
- Aria: warm, clear narrator
- Luna: soft, storytelling
Clear Instructors for Lessons
Sage delivers a calm, instructional tone perfect for grammar explanations or vocabulary drills. Its neutral delivery ensures instructions are understood without distraction.
Atlas also works well for news-style updates or structured lesson content, providing a professional backdrop for educational material.
- Sage: calm, instructional
- Atlas: neutral, news
Energetic Voices for Quizzes
Felix brings an energetic, promo-like energy to quiz interfaces, keeping users alert during rapid-fire vocabulary checks. Milo offers a friendly, conversational tone that feels less robotic during interactive dialogues.
These voices add dynamism to repetitive exercises, helping maintain user motivation during high-frequency practice modes.
- Felix: energetic, promo
- Milo: friendly, conversational
- Nova: bright, upbeat
Support for 23 Global Languages
Support your learners with high-quality synthetic voices in 23 languages, covering major European tongues and key Asian markets. Our API handles diverse linguistic requirements for flashcards and reading assistants without complex infrastructure.
Major European Languages
Build immersive exercises in Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swedish, and Turkish. Every voice can speak any of these languages, ensuring consistent audio quality across your curriculum.
Asian Languages with Latin Script
We support Malay, Turkish, and Swahili, expanding your reach beyond traditional Western markets. These languages allow you to create content for growing learner bases without needing separate voice models or specialized handling for non-Latin scripts.
Auto-Detection and Manual Override
If you do not set the language, the system guesses it from the script (defaulting to English for Latin text). For non-English Latin-script text, explicitly set the language parameter to ensure accurate pronunciation and avoid misinterpretation by the engine.
Simple Integration for Indie Developers
Integrating VerbaSound into your language learning app is straightforward because our API mirrors the familiar OpenAI speech endpoint structure. You can use existing OpenAI SDKs by simply swapping the base_url to https://api.kt10.info/v1 and updating your authentication key. This compatibility means less code rewriting and faster deployment for indie developers already comfortable with standard TTS workflows.
OpenAI SDK Compatible
Our API endpoint POST https://api.kt10.info/v1/audio/speech accepts the same JSON body shape as the OpenAI speech endpoint. This allows you to leverage official OpenAI SDKs with minimal configuration changes, reducing the learning curve for your development team.
- Use standard OpenAI SDKs by changing
base_url - Compatible with existing TTS integrations
- Supports
tts-1andtts-1-hdmodel strings
Standard JSON Requests
Send a simple JSON payload containing your text, chosen voice, and desired audio format. The API handles the synthesis and returns the audio file directly in the response. You can also specify language codes or let the system auto-detect from the script, ensuring correct pronunciation for diverse vocabulary lists.
- JSON body with
input,voice, andresponse_format - Optional
languageparameter for non-English Latin text - Returns audio directly in the HTTP response
Fast Response Times
Requests are processed sequentially, generating audio at roughly real-time speeds (25 characters per second). While short phrases may take a few seconds, longer texts are split by the client and joined by the client. Under load, requests may queue briefly, but typical usage remains snappy enough for interactive exercises and flashcard reviews.
- Roughly 25 characters generated per second
- Short phrases take a few seconds
- Requests processed sequentially with queuing under load
Developer-Friendly Pricing
You only pay for what you generate: $10 per 1 million characters, or $0.01 per 1,000 characters. Failed requests are free, so you never waste credits on errors. Your balance never expires, giving you flexibility for unpredictable student usage.
Pay Per Character
Spaces and punctuation count as characters, making cost calculation straightforward for vocabulary lists and short stories. This model eliminates the guesswork of flat-rate plans, allowing indie developers to scale audio production directly with user engagement.
- Cost: $0.01 per 1,000 characters
No Monthly Fees
There are no subscriptions, seats, or minimum spend requirements. You can pause usage and resume without losing access or incurring downtime charges. This structure is ideal for hobby projects or apps with seasonal traffic spikes.
Crypto Top-Up Support
Top up your balance instantly with USDT on the TRON network (TRC-20). Minimum deposit is $5, and credits appear automatically within minutes. We do not process cards, PayPal, or VAT invoices, keeping overhead low for small teams.
- Network: TRON (TRC-20) only
- Minimum top-up: $5
- Crediting time: automatic, within minutes

Technical Reliability for Education
Reliable audio output is essential for language learners who depend on clear pronunciation models. Our API delivers consistent synthetic speech across all 23 supported languages, ensuring that vocabulary flashcards and reading assistants sound uniform. Whether generating short phrases or long texts, the audio quality remains stable, helping students build accurate auditory memory without distracting variations.
Unlike some services, we do not use your input text to train our models. Your lesson content remains yours, used solely for immediate synthesis. Request metadata is retained only for billing purposes, keeping your educational data private and secure without adding unnecessary complexity to your app's infrastructure.
Consistent Audio Quality
We generate synthetic speech that maintains a consistent tone and pacing, which is critical for language acquisition. By using licensed recordings from LibriTTS-R, we ensure that every voice is distinct. This consistency helps learners distinguish between different characters in interactive stories or follow along with instructional lessons.
Each generated file includes an inaudible watermark, marking it clearly as synthetic speech. This helps maintain transparency in your educational tools, allowing you to confidently label audio assets as AI-generated without compromising the learning experience. The voices are designed to be clear and easy to understand, even for beginners.
No Data Training
Your input text is used exclusively to produce the audio file and is not stored for training purposes. This means that the specific vocabulary you teach or the stories you narrate remain your intellectual property. We do not mine your content to improve our models, ensuring a clean separation between your data and our engine.
Request metadata, such as timestamp and character count, is kept only for accurate billing. This minimal data retention policy simplifies compliance and gives you peace of mind when handling student data or sensitive educational materials in your language learning application.
Standard Audio Formats
We support mp3, wav, opus, flac, and pcm formats, giving you flexibility in how you deliver audio to your users. MP3 is ideal for web-based apps due to its small file size, while WAV offers high fidelity for offline educational tools. FLAC provides lossless quality, and PCM is perfect for raw audio processing.
Using standard formats ensures compatibility with existing media players and libraries in your stack. You can easily integrate these audio files into your flashcard systems, reading assistants, or quiz modules without needing custom codecs. This reduces development time and allows you to focus on creating engaging learning experiences.
Perfect for Flashcards and Reading
Language learning apps thrive on repetition and immersion, but maintaining consistent audio assets can be tedious. Our TTS API allows you to generate pronunciation for any vocabulary list on the fly, ensuring every card sounds natural and accurate without managing large audio libraries.
Text-to-Speech for Vocabulary
Generate instant audio for flashcards using simple JSON requests. With support for 23 languages and auto-detection, you can create dynamic exercises where users hear words pronounced in their native or target dialect, enhancing retention through auditory learning.
- Support for 23 languages including Arabic, Chinese, and Swahili
- Auto-detect language from script or force specific ISO codes
- Up to 4096 characters per request for rich vocabulary context
Interactive Story Modes
Turn static reading exercises into engaging audio stories. Send text to our API and receive high-quality synthetic voices that narrate lessons, keeping learners engaged with clear, consistent narration that adapts to the pace of the story.
Pronunciation Practice
Enable users to hear exact pronunciations for difficult words. Use the speed parameter to slow down speech for clarity or speed it up for natural flow, helping learners distinguish subtle phonetic differences in languages like Japanese or Korean.
- Adjustable speed from 0.25x to 4.0x
- Clear narration styles like 'sage' for instruction
- Inaudible watermark ensures synthetic origin is known
Get Started with Your API Key
Create your API key in seconds by visiting the /get-api-key/ page. Choose Google for instant sign-up or register with an email and password to start building your language learning tools.
Google Sign-Up Option
Sign in with Google to receive a free trial credit of $0.50 immediately. This credit covers about 50,000 characters and is valid for 7 days, giving you time to test the API without entering payment details.
Email Registration
Register with an email and password if you prefer not to use Google. Note that email-based accounts start without trial credit, so you will need to top up to begin generating audio.
Instant Access
Your account dashboard displays your API key, current balance, and usage statistics. You can use the provided curl example to make your first request immediately. Top up with USDT (TRC-20) for ongoing usage.
- View your active API key and balance
- Use the ready-made curl example
- Top up via USDT on TRON
- Issue a new key anytime if needed
Explore Our Voice Library
Explore our eight distinct voice presets, each crafted to support different learning contexts. From warm narrators to clear instructors, our library provides the auditory variety needed to keep language learners engaged.
Female Voice Options
Aria offers a warm, clear narration style ideal for storytelling. Nova brings a bright, upbeat energy perfect for quick drills. Luna provides a soft, storytelling tone for immersive reading, while Sage delivers a calm, instructional voice for structured lessons.
- Aria: Warm, clear narrator
- Nova: Bright, upbeat
- Luna: Soft, storytelling
- Sage: Calm, instructional
Male Voice Options
- Felix: Energetic, promo
- Milo: Friendly, conversational
- Atlas: Neutral, news
- Orion: Deep, documentary
Voice Character Descriptions
All voices are based on licensed recordings from LibriTTS-R and LibriVox. Each voice supports all 23 languages, allowing you to maintain consistent character identity across diverse linguistic exercises.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Per-character billing is the most predictable model for variable usage.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- LibriTTS-R paper (arXiv 2305.18802)
- Node.js documentation
- W3C WAI: making audio and video accessible
- ISO 639 language codes
- Python documentation
- LibriVox public-domain audiobooks
- IETF RFC 7845: Ogg Opus
- openai-python on GitHub
- Creative Commons Attribution 4.0
- TRON developer hub: TRC-20
- OpenAI API reference: create speech
