Frequently asked questions
What is the best voice for Arabic content?
Q: Which voice is best for Arabic content? A: All eight voices—aria, nova, luna, sage, felix, milo, atlas, and orion—support Arabic. You can choose any preset based on your preferred tone, such as the warm narrator style of aria or the neutral news delivery of atlas. Each voice is licensed from LibriTTS-R recordings and performs equally well across all supported languages without additional fees.
How fast is the Arabic text-to-speech generation?
Generation speed is roughly real-time, processing about 25 characters of text per second of audio. Short phrases may take a few seconds, while longer texts are processed sequentially. Under load, requests may wait in a queue, but typical latency remains consistent with the character count and processing queue status at the time of the request.
Does the API support other languages besides Arabic?
Yes, the API supports 23 languages including 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 you should set the language parameter explicitly for non-English Latin-script text to ensure correct pronunciation.
Can I use the generated audio for commercial streams?
The generated audio files are marked with an inaudible watermark indicating synthetic speech. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. You can use the output for streams.
How much does it cost to generate 10,000 characters?
The cost is $0.10 for 10,000 characters. The pricing is $10 per 1 million characters, which equals $0.01 per 1,000 characters. Spaces and punctuation count toward the character total. Failed requests are free, so you only pay for successfully synthesized audio files delivered in the response.
Is there a free trial available?
Yes, signing in with Google provides a $0.50 credit valid for 7 days, allowing approximately 50,000 characters. Accounts created with email and password do not receive this trial credit. The balance does not expire, and you can top up with USDT on the TRON network for ongoing usage without monthly plans or minimum spends.
What audio formats are supported?
The API supports mp3, wav, opus, flac, and pcm formats. You specify the desired format using the response_format field in your JSON request body. All formats are generated using the same engine and pricing model, allowing you to choose the balance between file size and quality for your specific streaming or development needs.
Do I need to pay for each voice type?
No, all eight voices use the same engine and price. Whether you choose tts-1 or tts-1-hd, the cost per character remains identical. You can switch between voices like aria, nova, or atlas without any additional fees, as the pricing is based solely on the input character count and not the voice or model selected.
How do I handle texts longer than 4096 characters?
You must split the text into parts of up to 4096 characters each and send separate requests. Then, join the resulting audio files. The API does not support streaming or real-time partial audio, so you handle the concatenation in your application logic. This ensures each request stays within the character limit for successful synthesis.
What happens if a request fails?
If synthesis fails, the charge is returned in full. You are not billed for failed requests. The API returns the status, allowing you to retry or handle the error in your workflow. This ensures you only pay for successfully generated audio files, maintaining cost efficiency even during high-load periods or temporary service issues.
Is there a limit on how many voices I can use?
There is no limit on the number of voices you can use. You can switch between any of the eight presets—aria, nova, luna, sage, felix, milo, atlas, or orion—across your requests. Each voice supports all 23 languages, allowing you to mix and match voices freely without additional costs or restrictions based on your account type.
Can I change my API key if it gets compromised?
Yes, you can issue a new key at any time. When you generate a new key, the old one stops working immediately. You can manage your keys and view your balance on the /get-api-key/ page. This allows you to quickly revoke access if a key is compromised without affecting your existing balance or usage history.
Does the API support SSML for better control?
No, the API does not support SSML. You cannot use phoneme dictionaries or custom pronunciation lexicons. Control is limited to the speed parameter (0.25–4.0) and the voice selection. For precise timing or pronunciation, you must manage these aspects in your application logic or by splitting text into smaller, manageable chunks before sending requests.
Is there a mobile app for this API?
No, there is no official mobile app or mobile SDK. You can use the API via the OpenAI SDK or plain HTTP requests from any device. The endpoint is POST https://api.ktvsuper.info/v1/audio/speech, and you can integrate it into your mobile applications by sending JSON bodies with your API key in the authorization header.
First request
curl
curl https://api.ktvsuper.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 Arabic TTS Matters for Streamers
Arabic represents a massive, underserved audience for VTubers and indie streamers. By localizing your content, you unlock direct engagement with millions of viewers who prefer their native language, significantly expanding your reach beyond English-only communities.
Consistency is key for character immersion. Using a dedicated TTS API ensures your Arabic character voice remains stable and recognizable across long streaming sessions, regardless of the duration or time of day.
Expanding your audience reach with localized content
Localization goes beyond simple translation; it requires natural speech patterns. Our API supports Modern Standard Arabic, allowing you to produce high-quality audio that resonates with a broad Arab-speaking demographic without needing external voice actors.
Consistency in character voice across long streams
Unlike human voice actors who may vary in tone or energy, a synthetic voice maintains exact consistency. This is crucial for VTuber branding, where your character's voice is a core asset. Whether it is a 10-minute update or a 24-hour stream, the voice remains identical.
Reducing vocal fatigue for 24/7 streamers
Streaming for hours on end causes vocal strain. Offloading Arabic dialogue to TTS preserves your natural voice for interactive segments. This allows you to maintain high energy levels throughout long streams without the physical toll of continuous speaking.
Voice Catalog Overview
EchoWave provides eight distinct synthetic voices, each engineered from licensed LibriTTS-R recordings. This foundation ensures consistent, high-quality audio suitable for character-driven VTuber content without the need for custom voice cloning or extensive training data.
Female voices: Aria, Nova, Luna, Sage
Aria offers a warm, clear narrator tone ideal for exposition. Nova delivers a bright, upbeat energy perfect for energetic segments. Luna provides a soft, storytelling quality, while Sage presents a calm, instructional demeanor for guides or tutorials.
- Aria: warm, clear narrator
- Nova: bright, upbeat
- Luna: soft, storytelling
- Sage: calm, instructional
Male voices: Felix, Milo, Atlas, Orion
Felix brings an energetic, promo-ready voice to announcements. Milo offers a friendly, conversational style for direct address. Atlas provides a neutral, news-like tone, and Orion delivers a deep, documentary-style voice for immersive backgrounds.
- Felix: energetic, promo
- Milo: friendly, conversational
- Atlas: neutral, news
- Orion: deep, documentary
Arabic language support and voice selection
All eight voices support Arabic, allowing you to select the personality that best fits your character rather than being limited by language-specific voice assignments. Set the language field to ar for accurate pronunciation.
- All voices support Arabic
- Set
language: ar - Consistent voice across languages
Technical Performance Metrics
Synthesis speed averages ~25 characters per second of generated audio, effectively matching real-time playback for long texts. Short phrases typically require a few seconds to process, ensuring immediate feedback for quick voice checks.
Generation speed: ~25 characters per second
Text-to-speech processing maintains a consistent throughput of roughly 25 characters per second. This ratio allows streamers to predict audio duration accurately, facilitating precise scheduling for VTuber avatar lip-syncing during extended broadcasts.
Latency expectations for short vs long texts
Short inputs under 4096 characters return quickly, though initial processing may take a few seconds. Longer texts are split into chunks; while each chunk generates in real-time, the cumulative wait depends on queue depth.
Queue handling during high load periods
Requests are processed sequentially. Under high load, synthesis jobs may wait in a queue before beginning. This linear processing ensures stability but can increase latency during peak usage times for indie developers.
Pricing Structure for Indie Developers
EchoWave offers a transparent, usage-based pricing model designed for indie developers and streamers who need predictable costs without monthly commitments. The standard rate is $10 per 1 million characters, which equals $0.01 per 1,000 characters. Spaces and punctuation count toward this total, ensuring accurate billing for every byte of input text processed by the TTS engine.
Cost per million characters ($10)
The API charges strictly on character volume. Failed requests are free, with the charge returned in full if synthesis fails. There are no monthly plans, seats, or minimum spend requirements, allowing you to scale your VTuber output based on actual usage rather than fixed overhead.
Free trial credit with Google sign-in
Accounts created via Google sign-in receive $0.50 of trial credit, providing approximately 50,000 characters of generation capacity. This credit is valid for 7 days, allowing you to test voice quality and integration workflows immediately. Note that accounts created with email and password start without this trial credit.
USDT (TRC-20) top-up options
For ongoing needs, top-ups are processed via USDT on the TRON network (TRC-20). The minimum top-up amount is $5, and funds are credited automatically within minutes of transfer. The platform does not accept credit cards, PayPal, or offer subscriptions. Your balance never expires, and no VAT invoices are generated.
Audio Format Options
EchoWave supports MP3, WAV, Opus, FLAC, and PCM formats via the response_format parameter, giving developers control over file size and quality. Each generated audio file carries an inaudible watermark to mark it as synthetic speech, ensuring transparency for your audience.
MP3 for efficient streaming playback
MP3 is an efficient format for streaming playback. It offers a good balance between file size and audio fidelity, making it ideal for low-latency delivery to viewers. This format minimizes bandwidth usage during live streams without sacrificing intelligibility.
- Optimized for live streaming workflows
- Smallest file size among supported lossy formats
- Wide compatibility with streaming platforms
WAV for high-quality archival
WAV provides uncompressed audio, preserving the full fidelity of the generated voice. This format is suitable for high-quality archival of stream highlights or creating pre-recorded segments where bandwidth constraints are less critical. It ensures the audio remains pristine for editing.
- Uncompressed PCM audio data
- Ideal for post-production editing
- Maintains original voice nuances
FLAC and Opus for lossless needs
FLAC offers lossless compression, reducing file size while retaining all audio information. Opus is another efficient lossy format, excellent for voice communication. Both options allow developers to choose between strict lossless preservation or highly efficient compression for specific use cases.
- FLAC: Lossless compression for perfect quality
- Opus: Efficient for voice-optimized transmission
- PCM: Raw data for maximum compatibility

Integration for Developers
Integrating EchoWave into your VTuber pipeline is straightforward because the API mirrors the OpenAI speech endpoint structure. You can use official OpenAI SDKs by simply pointing them to our base URL, ensuring familiar workflows for indie developers already accustomed to their syntax.
This compatibility means you can swap providers or test alternatives with minimal code changes. The consistent request shape allows you to focus on voice selection and stream timing rather than learning new authentication or payload patterns.
OpenAI SDK compatibility
The API accepts the same JSON body as the OpenAI speech endpoint. By updating the base_url to https://api.ktvsuper.info/v1, existing OpenAI SDK clients work immediately. This includes support for standard parameters like model, voice, and response_format without requiring custom wrappers or proprietary client libraries.
- Uses standard OpenAI request shapes
- Works with official OpenAI SDKs
- No custom client library needed
Simple cURL requests for quick tests
For rapid validation, you can send a POST request to https://api.ktvsuper.info/v1/audio/speech with your tts-... bearer token. Pass the text in the input field and specify your preferred voice and response_format. The endpoint returns the audio file directly, allowing you to pipe it straight into your broadcasting software or save it for later use.
Base URL configuration for API access
All requests should target https://api.ktvsuper.info/v1 as the base URL. Ensure your Authorization header includes Bearer tts-... with your unique key. The endpoint POST /v1/audio/speech handles synthesis, while GET /v1/voices allows you to programmatically list available presets and languages before initiating generation.
Voice Licensing and Ownership
EchoWave voices are built on licensed recordings from the LibriTTS-R dataset (CC BY 4.0), sourced from LibriVox public-domain audiobook readings. This foundation ensures that the synthetic voices used in your stream content rest on clear, open licensing terms.
LibriTTS-R and CC BY 4.0 licensing
The underlying voice models derive from the LibriTTS-R dataset, which operates under the Creative Commons Attribution 4.0 International license. This means the base recordings are freely available for reuse, provided proper attribution is given to the original dataset contributors.
LibriVox public-domain source material
Source audio originates from LibriVox, a platform dedicated to public-domain literature. Because the original readings are in the public domain, they are free from complex copyright restrictions, allowing EchoWave to process and synthesize them into stable, reusable voice presets without ongoing royalty obligations.
Commercial use rights for streamers
Streamers and indie developers can use generated audio commercially in their broadcasts and VODs without additional licensing fees. The synthetic nature of the output, combined with the public-domain base, simplifies rights management for content creators who need reliable, clear voice assets for character work.
- Commercial use included in standard API pricing
- Public-domain source material reduces legal complexity
- No ongoing royalty payments per stream
Data Privacy and Storage
EchoWave prioritizes your data integrity by ensuring that the input text you submit for synthesis is used exclusively for generating the audio file. It is not retained, analyzed, or added to any training datasets for model improvement, keeping your unique character lines or stream scripts private.
Input text not used for training models
Unlike many competitors, we do not repurpose your submitted text to refine our voice engines. Your scripts remain yours, with no secondary use for AI training or data aggregation.
Generated audio returned instantly
The synthesized audio file is delivered directly in the API response and is not stored on our servers after delivery. This ensures that your character voices or stream assets are transient and secure, reducing the risk of unauthorized access to your content.
Metadata retention for billing purposes
We retain minimal request metadata, including the timestamp, character count, and status, solely for accurate billing and troubleshooting. This data is used to verify usage against your account balance and does not include the content of your text or audio.
Support and Limitations
We prioritize development over customer service overhead, so all inquiries are handled via email. There is no phone number or live chat option available for streamers or developers needing quick clarifications on API behavior or billing issues.
As an independent API provider, we do not offer formal SLAs or uptime guarantees. While we strive for reliability, our infrastructure is designed for cost-efficiency rather than enterprise-grade redundancy, meaning occasional downtime may occur without compensation.
Email-only support channel
Our support team responds to emails as capacity allows. We do not have live chat or phone support lines. For urgent issues during a stream, check the documentation first, as most common integration errors are documented.
No SLA or uptime guarantees
We do not publish Service Level Agreements or guarantee specific uptime percentages. Our infrastructure is optimized for indie developers rather than enterprise clients requiring strict contractual reliability. You should build retry logic into your application.
Rate limits: 60 requests per minute
Each API key is limited to 60 requests per minute. New accounts start with a lower limit of 10 requests per minute until the first top-up is processed. Exceeding this limit returns a 429 error.
- Standard limit: 60 requests/minute per key
- Initial limit: 10 requests/minute
- Reset: Automatic, sliding window
- Error code: 429 Too Many Requests
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Each request is processed individually; under load, requests may queue, taking a few seconds for short phrases or roughly real-time for longer texts.
23 languages
Any voice speaks any of these languages.
