Frequently asked questions
Is the voice generated by a human?
No, the audio is synthesized by an API. The voices are based on licensed recordings from LibriTTS-R and LibriVox public-domain audiobook readings. Each generated file carries an inaudible watermark marking it as synthetic speech, ensuring you know exactly what you are using for your course modules.
Can I control the speed of the voice?
Yes, you can adjust the speed between 0.25 and 4.0. This allows you to fine-tune narration for your online modules, ensuring the pacing matches your instructional design needs. Simply include the speed parameter in your JSON request body to the API endpoint.
Do the voices sound robotic?
The voices are based on licensed recordings from LibriVox public-domain audiobook readings, offering clear, natural tones. While they are synthetic, they avoid the typical robotic sound of older TTS systems. They are designed for clear, character-based voice lines suitable for educational content.
Can I use the voices for commercial courses?
Yes, you can use the generated audio for commercial online courses. The API charges based on characters processed, with failed requests being free. There are no restrictions on usage within your LMS or course materials, making it ideal for scalable educational content production.
How accurate is the pronunciation?
The API supports 23 languages, including English, Spanish, and French. If the language is not set, it guesses from the script, defaulting to English for Latin text. For non-English Latin-script text, you should explicitly set the language parameter to ensure accurate pronunciation.
Can I mix different voices in one audio?
The API generates one audio file per request. To mix voices, you generate separate audio files for each voice and join them using your preferred audio editing tool. This gives you full control over the final composition of your course audio tracks.
Do I need to record my own voice?
No, you do not need to record your own voice. You select from eight presets like Aria, Nova, or Felix. This saves time and money compared to hiring voice actors, allowing you to generate character dialogues and narrator tones directly via the API.
Is the audio format MP3 only?
No, you can choose from mp3, wav, opus, flac, or pcm. This flexibility allows you to select the format that best fits your LMS requirements. For example, use WAV for high-quality editing or MP3 for efficient web delivery in your online courses.
Can I change the voice after generation?
No, the voice is fixed when the audio is generated. To change the voice, you must send a new request with the desired voice parameter. Since requests are processed sequentially, you can quickly generate alternative versions to find the best fit for your module.
Does the API work offline?
No, this is a cloud API. You need an internet connection to send requests to https://api.combowin.top/v1/audio/speech. The API processes requests and returns the audio file, so all generation happens on their servers, not on your local device or LMS.
Can I use the voices in video editing?
Yes, you can download the generated audio files and import them into your video editing software. The API returns the audio file in the response, which you can then sync with your video content. This is a common workflow for creating engaging online course videos.
Is there a limit to how many voices I can use?
You can use any of the 8 available voices in your requests. There is no limit to the number of different voices you can use across your projects. You simply specify the voice parameter in each request to switch between characters like Luna or Atlas.
How do I choose the right voice for my course?
Choose based on the tone you need: Aria for clear narration, Felix for energetic promos, or Atlas for neutral news. Use the /v1/voices endpoint to list all options. Test different voices with short texts to see which best fits your instructional style and audience.
Can I get a sample of each voice?
Yes, use the GET /v1/voices endpoint to retrieve information about all voices, including their descriptions. You can then generate short audio samples by sending requests with each voice preset. This helps you evaluate which voice works best for your specific course content.
Is the API available 24/7?
The API is a cloud service available whenever you need it. There is no SLA or uptime guarantee. Requests are processed sequentially, so under load, you might experience slight delays, but the service remains available.
Can I integrate this with WordPress?
There is no official WordPress plugin or browser extension. However, you can integrate it by using the OpenAI SDK or plain HTTP requests in your custom WordPress theme or plugin. You just need to change the base URL to the API endpoint and use your key.
Definition of synthetic host voices
Synthetic host voices transform written text into spoken audio using cloud-based TTS APIs. Instead of recording human actors, educators input scripts directly into the system. The technology converts characters into natural-sounding speech, enabling rapid production of quiz introductions and feedback lines for online modules.
Text-to-speech conversion
The API accepts text input and returns audio files. You send a POST request with your script, and the engine generates the speech. This eliminates the need for recording studios or scheduling talent, allowing you to generate voice lines on demand.
Character-based voice presets
AudioTeach offers eight distinct voices, including energetic options like Felix for promo-style segments or calm tones like Sage for instructional content. These presets act as consistent character anchors, ensuring your game show host sounds uniform across all lessons without re-recording.
Automated audio production
Once you define the text and select a voice, the system handles the synthesis. You receive an audio file ready for embedding in your LMS. The process is fully automated, letting you scale content creation without managing voice actor contracts.
Benefits for online course creators
Educators often need distinct characters—like a quizmaster or narrator—to keep learners engaged without hiring multiple voice actors. AudioTeach provides eight preset voices, each with a specific tone, allowing you to maintain consistent character presence across all your modules.
Consistent character presence
Each voice preset, such as Felix for energetic promos or Luna for storytelling, delivers a uniform tone across every lesson. This consistency helps students recognize characters instantly, building a cohesive narrative without the variability of different human actors.
Scalable content creation
Generate thousands of voice lines by simply pasting text into the API. With a limit of 4,096 characters per request, you can scale your course library rapidly, creating new quizzes or feedback loops without waiting for studio schedules (splitting longer texts into parts).
Cost reduction vs voice actors
At $0.01 per 1,000 characters, generating audio is significantly cheaper than booking professional voice talent. Failed requests are free, and you only pay for what you use, making it ideal for budget-conscious instructional designers.
How TTS engines process text
To generate a game show host voice, the API first receives your text via a POST request. The input is analyzed for structure and length, ensuring it stays within the 4,096-character limit per request. This step ensures the text is ready for synthesis without formatting errors.
Input text analysis
The engine parses the submitted string, checking for valid characters and language cues. If no language is specified, it guesses based on the script, defaulting to English for Latin text. This preprocessing ensures the voice model applies the correct pronunciation rules immediately.
- Validates input length against the 4,096-character limit
- Detects language script for accurate pronunciation
- Prepares text for the synthesis engine
Voice synthesis engine
The core engine applies the selected preset, such as Felix for an energetic host or Luna for a softer narrator. It processes the text using licensed recordings from LibriVox, generating speech that matches the chosen persona’s tone and pacing without needing manual direction.
- Applies specific voice characteristics (e.g., Felix for promo)
- Uses licensed LibriTTS-R source recordings
- Maintains consistent character presence across segments
Audio file generation
Once synthesized, the audio is encoded into your chosen format, such as MP3 or WAV. The system adds an inaudible watermark to mark it as synthetic. The file is then returned instantly via the API response, ready for integration into your LMS player.
- Supports MP3, WAV, OPUS, FLAC, and PCM
- Adds inaudible synthetic audio watermark
- Returns file directly in the API response
Role of game show hosts in education
Game show host voices bring dynamic energy to educational content, transforming static quizzes into interactive experiences. By using presets like Felix for energetic prompts or Luna for storytelling, educators can create consistent character presence across modules without hiring voice actors.
Quiz introductions
Host voices set the tone for assessments, making them feel like engaging challenges rather than tests. An energetic voice like Felix can introduce multiple-choice questions with enthusiasm, while a calm voice like Atlas provides neutral, clear instructions for serious evaluations.
- Create consistent opening hooks for every quiz module
- Differentiate between casual practice and formal assessment tones
- Reduce cognitive load by framing questions clearly
Feedback delivery
Synthetic voices provide immediate, consistent feedback for correct or incorrect answers. Unlike human actors who may vary in tone, a preset like Sage ensures instructional clarity, while Milo offers a friendly conversational style that encourages learners to try again without feeling criticized.
- Maintain encouraging tone even during repetitive feedback loops
- Scale personalized responses without increasing production time
- Ensure consistent auditory branding across all lesson modules
Engagement enhancement
Dynamic voice styles keep learners attentive by mimicking the excitement of live game shows. The ability to switch between warm narrators and upbeat hosts helps maintain pacing in long courses, preventing monotony and sustaining interest through varied auditory stimuli.
- Use voice variety to signal topic transitions
- Leverage energetic presets for high-stakes knowledge checks
- Create memorable audio signatures for course sections
Technical requirements for integration
Connecting to the AudioTeach API is straightforward: send a POST request to https://api.combowin.top/v1/audio/speech with your Bearer token and JSON payload. The endpoint accepts standard fields like model, input text, and voice preset, returning an audio file directly in the response.
For LMS integration, ensure your platform can handle the generated audio formats. The API supports mp3, wav, opus, flac, and pcm, ensuring compatibility with most modern educational platforms and web-based audio players without requiring complex conversion steps.
API endpoint connection
Authentication uses a Bearer token issued upon account creation. Simply set the Authorization header to Bearer tts-... and send your text payload. The system processes requests sequentially, so longer scripts may queue slightly during high usage.
- Endpoint: POST https://api.combowin.top/v1/audio/speech
- Auth: Bearer token in Authorization header
- Rate limit: 60 requests/minute (10 before first top-up)
Audio format compatibility
Choose from mp3, wav, opus, flac, or pcm based on your LMS requirements. MP3 is recommended for web delivery due to small file sizes, while WAV offers lossless quality for archival purposes.
- Supported formats: mp3, wav, opus, flac, pcm
- Character limit: 4096 per request
- Speed adjustment: 0.25–4.0x
LMS audio player support
Most modern Learning Management Systems support standard HTML5 audio tags, making integration seamless. Simply embed the returned audio URL or file into your course module. Ensure your player supports the chosen format for smooth playback across devices.
Voice characteristics and styles
Voice selection defines the personality of your educational modules. By choosing between energetic and calm tones, you can match the audio to the specific context of the lesson, whether it’s a high-energy quiz or a reflective summary.
Energetic vs calm tones
Use Felix for promo-style introductions or Milo for friendly, conversational feedback. Atlas provides a neutral, news-like tone for straightforward instructions, while Orion’s deep voice suits documentary-style explanations. This range allows you to maintain consistent character presence without hiring multiple actors.
- Felix: Energetic, ideal for promos
- Milo: Friendly, conversational tone
- Atlas: Neutral, news-style clarity
- Orion: Deep, documentary feel
Narrator vs conversational styles
Aria offers a warm, clear narrator style perfect for structured lessons. Luna’s soft voice suits storytelling modules, while Sage provides a calm, instructional tone for complex concepts. Nova adds a bright, upbeat energy for engaging quizzes or interactive elements.
- Aria: Warm, clear narrator
- Luna: Soft, storytelling focus
- Sage: Calm, instructional clarity
- Nova: Bright, upbeat energy
Age and gender variety
With eight distinct presets, you can create diverse cast members for your modules. The voices are based on licensed recordings from LibriTTS-R, ensuring natural prosody and realistic delivery. This variety helps keep learners engaged by providing auditory variety across different characters in your course.

Language support and localization
Language support is central to creating accessible online modules. Our API handles 23 languages, allowing you to localize course content for global audiences without hiring separate voice talent for each region.
You can generate consistent character voices across different linguistic contexts, ensuring your game show host or narrator remains recognizable whether teaching English or Spanish.
23 supported languages
We support Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. Every voice preset can speak any of these languages, giving you flexibility in casting your educational characters.
This wide range allows you to create diverse learning experiences without technical constraints. You are not limited to English-only modules if your target audience speaks another supported language.
- Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Auto-detection features
If you do not specify a language, the system guesses it from the script. Latin script text defaults to English, which is convenient for simple English lessons.
For non-English Latin-script text, auto-detection may not be accurate, so manual setting is recommended for precision.
Manual language setting
Set the language parameter to an ISO code to ensure correct pronunciation. This is crucial when using mixed scripts or when the default guess is wrong.
This control ensures your game show host sounds authentic in the target language, maintaining engagement in your online modules.
Quality and realism of synthetic voices
The synthetic voices in our API are built from licensed source recordings, specifically drawn from the LibriTTS-R dataset and public-domain audiobook readings from LibriVox. This foundation ensures that the generated audio retains a natural, human-like quality without sounding artificially robotic or disjointed.
Licensed source recordings
By using high-quality, pre-licensed recordings as the base for our voice presets, we ensure that each character—from the energetic host to the calm narrator—has a distinct, consistent personality. This approach allows educators to create engaging course modules with voice lines that feel authentic and professional, eliminating the need for expensive studio sessions or custom voice actor hiring.
Watermark identification
Every audio file generated by our API includes an inaudible watermark that clearly marks the content as synthetic speech. This subtle feature helps maintain transparency, allowing learners and institutions to easily distinguish between human-recorded and AI-generated audio elements within your online modules or interactive lessons.
Natural prosody
Our TTS engine processes text to produce natural prosody, mimicking the rhythm, stress, and intonation of human speech. This results in clear, character-based voice lines that enhance engagement without requiring complex SSML markup (which is not supported). The voices deliver feedback and quiz introductions with a tone that feels conversational and appropriate for educational contexts.
Use cases in different industries
AudioTeach’s API enables diverse educational sectors to replace expensive voice actors with consistent, character-based audio lines. By leveraging eight distinct presets, developers can assign specific personalities to different learning segments without hiring talent.
This approach ensures uniform quality across modules, whether producing a formal corporate briefing or a playful quiz introduction (though no SLA is provided). The system handles character limits by splitting longer scripts, maintaining narrative flow.
Corporate training modules
Compliance and onboarding modules benefit from clear, instructional voices like Sage or Atlas. These presets deliver neutral, professional tones that reduce listener fatigue during lengthy policy explanations.
- Consistent narrator tone across hundreds of videos
- Easy updates when policies change
- No need to re-record with human actors
K-12 interactive lessons
Educators can assign energetic voices like Felix for quiz hosts or Luna for storytelling segments. This variety keeps young learners engaged without the variability of different human narrators.
- Distinct characters for different lesson types
- Quick generation of new lesson audio
- Support for multiple languages in classrooms
Language learning apps
With support for 23 languages, developers can generate pronunciation examples in diverse tongues. The API allows manual language setting to ensure accuracy, even with similar scripts.
- Support for 23 languages including Arabic and Korean
- Manual language override for ambiguous text
- Consistent voice across different languages
Limitations of current TTS technology
Synthetic voices lack explicit emotion tags, so you cannot instruct a host to sound "excited" or "suspenseful" directly via the API. You must rely on the inherent tone of the preset—like Felix’s energetic promo style or Luna’s soft storytelling—to convey mood through word choice and pacing alone.
No emotion tags
Unlike human actors, the engine does not support emotional inflection commands. If your script requires a sudden shift in tone, you may need to adjust the script text or select a different voice preset that better matches the desired atmosphere.
No custom voice cloning
We do not offer custom voice cloning or uploading your own voice recordings. Each of the eight presets is based on licensed recordings from LibriTTS-R or LibriVox. You cannot create a unique brand voice or clone a specific instructor’s voice for your modules.
Character limit per request
Each API call accepts a maximum of 4096 characters. For longer game show segments or extended quiz introductions, you must split the text into smaller parts, generate each audio file separately, and join them in your LMS player. This adds a slight step to your workflow but keeps costs predictable.
Future of voice generation in e-learning
The evolution of TTS in e-learning will likely prioritize even more natural prosody, making character dialogues indistinguishable from human recordings. As models refine their intonation, educational modules can achieve higher engagement without the logistical overhead of recording sessions for every lesson update.
Processing speeds will continue to improve, reducing the wait time for long-text generation. This allows instructors to generate audio assets in near real-time, streamlining the workflow from script to published course content.
Improved realism
Future iterations will focus on nuanced intonation and pacing, allowing voices like Aria or Felix to convey subtle narrative cues. This enhances the immersive quality of game show styles and educational storytelling.
Refined audio processing will further blur the line between synthetic and human speech, making character-based learning experiences more engaging and emotionally resonant for students.
- Enhanced prosody for natural dialogue
- Subtle intonation variations
- Immersive character experiences
Faster processing speeds
As computational efficiency increases, the generation time for long texts will decrease significantly. This enables educators to produce audio content more rapidly, supporting agile course development.
Reduced latency will make it easier to integrate TTS into interactive learning modules, allowing for dynamic audio responses that keep students engaged.
- Quicker audio generation
- Support for agile development
- Dynamic audio integration
Broader language support
While we currently support 23 languages, future updates may expand this list to include more regional dialects and less common languages. This will make e-learning accessible to a global audience.
Better language support will enable educators to create localized content with high-quality voices, ensuring consistent user experience across different markets.
- Expanded language list
- Regional dialect support
- Global accessibility
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Technical terms and specific instructions are articulated with precision.
23 languages
Any voice speaks any of these languages.
Sources and further reading
- Wikipedia
- openai-python on GitHub
- FFmpeg documentation
- TRON developer hub: TRC-20
- LibriTTS-R paper (arXiv 2305.18802)
- Tether (USDT)
- LibriVox public-domain audiobooks
- LibriTTS-R corpus (OpenSLR 141)
- Python documentation
- W3C WAI: making audio and video accessible
- OpenAI API reference: create speech
- IETF RFC 7845: Ogg Opus
