Frequently asked questions
What text-to-audio Hindi voices are available?
We provide eight distinct voice presets: Aria, Nova, Luna, and Sage (female); Felix, Milo, Atlas, and Orion (male). Any voice can speak Hindi. You can specify the language in the request to ensure accurate pronunciation, though the system also attempts automatic detection based on the script provided in the input text.
Can I use these voices for commercial restaurant apps?
Yes. The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings. There are no restrictions on commercial usage for your restaurant platform. Generated audio files carry an inaudible watermark to identify them as synthetic speech, which is standard for this technology.
How do I integrate the API with my existing tech stack?
The API follows the same request shape as the OpenAI speech endpoint. You can use the official OpenAI SDKs by changing the base_url to https://api.svajpa.net/v1 and providing your API key. Alternatively, send a POST request to /v1/audio/speech with JSON body containing model, input, voice, and other parameters.
Is there a limit on characters per request?
Each API request accepts a maximum of 4096 characters for the input text field. For longer texts, such as full menus, you should split the text into parts of up to 4096 characters, generate audio for each part, and then join the resulting audio files in your application logic.
Do you support streaming audio for real-time menu reading?
No, we do not support streaming or WebSocket connections for real-time partial audio. The API returns the complete audio file upon request completion. There is no streaming or WebSocket support. Requests are processed sequentially, so under load, a request may wait in a queue before synthesis begins, meaning true real-time streaming is not available.
What happens if my API request fails?
If synthesis fails, the charge is returned in full. You are only billed for successful requests. Failed requests incur no cost, ensuring you do not pay for audio that was not successfully generated. You can retry the request with the same parameters after handling the error response.
Are the voices licensed or cloned from real people?
The voices are based on licensed recordings from LibriTTS-R (CC BY 4.0) and public-domain audiobook readings from LibriVox. They are not custom voice clones of specific individuals. Each generated file carries an inaudible watermark marking it as synthetic speech, distinguishing it from human-recorded audio.
Can I mix different voices in one audio file?
No, each API request generates audio for a single voice preset. To mix voices, you must make separate requests for each segment using different voice parameters, then join the resulting audio files in your application. The API does not support multi-voice generation within a single response.
How fast is the text-to-speech Hindi generation?
Typical speed is roughly real-time, generating about 25 characters of text per second of audio. Short phrases may take a few seconds. Requests are processed sequentially, so under load, a request may wait in a queue, potentially increasing total generation time for long menus.
Do I need to pay for every character including spaces?
Yes, the price is $10 per 1 million characters, which includes spaces and punctuation. You are charged based on the total character count of the input text sent in the request body. This applies to all successful requests, regardless of the language or voice used.
Is there a minimum spend to start using the API?
No, there is no minimum spend or monthly plan. You can start with a free trial of $0.50 credit (valid for 7 days) when signing in with Google. For top-ups, the minimum amount is $5 via USDT on the TRON network (TRC-20), credited automatically within minutes.
Can I download the voice samples before buying?
The API documentation lists available voices and their characteristics. You can generate sample audio by making a small request with your trial credit or after topping up. There is no dedicated sample download page; you must generate audio to hear the specific voice output for your text.
Do you offer voice cloning for specific brand voices?
No, we do not offer custom voice cloning or uploading your own voice. The catalog consists of eight fixed presets based on licensed recordings. You cannot design custom voices or clone specific individuals. If you need a unique voice, you must select the closest preset from our eight options.
How do I renew my API key for security?
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 without downtime. You can manage your active key and view usage details on the /get-api-key/ page.
First request
curl
curl https://api.svajpa.net/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 Voice Matters for Dining Apps
For restaurant platforms, voice transforms static menus into engaging experiences. Clear audio ensures readability, allowing diners to absorb dish descriptions without straining their eyes. This reduces friction in digital ordering flows and keeps users focused on the culinary content rather than parsing dense text blocks.
Accessibility compliance becomes straightforward when audio supplements visual interfaces. By offering spoken alternatives, dining apps cater to users with visual impairments or reading difficulties, ensuring broader inclusivity without requiring complex custom voice design. This simple addition respects user diversity while maintaining a clean, modern interface.
Menu Readability
Voice narration aids users who prefer listening over reading lengthy ingredient lists. It ensures that allergen warnings and dietary details are heard clearly, reducing order errors and improving satisfaction for those with visual strain or dyslexia.
- Enhances readability for complex menu items
- Reduces visual strain for diners
- Clarifies allergen and dietary details instantly
Accessibility Compliance
Providing audio options helps meet digital accessibility standards, such as WCAG guidelines, by offering non-text content alternatives. This demonstrates a commitment to inclusive design, making dining platforms usable for a wider audience, including those relying on screen readers or audio assistance.
- Supports WCAG guidelines for non-text content
- Caters to visual impairment needs
- Improves overall user inclusivity
Brand Tone Alignment
Voice sets the emotional tone of the dining experience. A warm narrator voice can make a menu feel inviting, while a crisp, clear voice might suit a fast-paced delivery app. Aligning voice with brand identity enhances user engagement without needing custom voice cloning or expensive production.
Indie developers can choose from preset voices that match their restaurant’s vibe, ensuring consistency across the platform. This alignment builds trust and makes the digital ordering process feel more personal and polished, enhancing the overall brand perception.
The Hindi Voice Catalog
Svajpa Editorial presents three female presets from our licensed catalog, each optimized for distinct dining app contexts. These voices are built on LibriTTS-R recordings, ensuring high-quality synthetic speech that enhances user engagement without custom design overhead. All presets support Hindi and 22 other languages, allowing you to maintain consistent brand tone across diverse menu sections.
Aria: Warm Narrator
Aria provides a warm, clear narration ideal for reading complex menus or describing dishes. She is one of four female presets. Her tone is particularly effective for detailed ingredient lists or allergen warnings, where clarity and warmth help users navigate information without feeling rushed.
- Ideal for detailed menu descriptions and allergen info
Nova: Bright Upbeat
Nova brings a bright, upbeat energy to your platform. She is one of four female presets. This voice works well for promotional banners, daily specials, or quick updates, adding a lively touch that encourages users to explore new offerings on your dining platform.
- Best for daily specials and promotions
Luna: Soft Storytelling
Luna offers a soft, storytelling voice perfect for culinary histories or chef profiles. She is one of four female presets. Her gentle tone invites users to linger, making it suitable for background narration in immersive dining experiences or detailed origin stories of dishes.
- Perfect for chef profiles and dish origins
Technical Simplicity for Indie Devs
Indie developers building restaurant apps can use the official OpenAI SDKs with zero friction. Simply point the base_url to https://api.svajpa.net/v1 and inject your tts-... key. This ensures your existing code works immediately without learning a proprietary client.
Requests follow a standard JSON shape identical to OpenAI’s speech endpoint. Send text via POST /v1/audio/speech, and receive audio files directly. No complex streaming protocols or custom wrappers are needed for basic text-to-audio conversion. The API returns complete audio files.
OpenAI SDK Compatibility
The API accepts the same request structure as OpenAI, allowing you to use familiar libraries. This reduces integration time for developers already comfortable with OpenAI’s ecosystem, ensuring a smooth transition for dining platform features.
Simple JSON Requests
Send a single JSON payload containing model, input, voice, and response_format. The system processes the text and returns the audio file. Failed requests incur no charge, simplifying error handling in your app.
Multiple Audio Formats
Choose between mp3, wav, opus, flac, or pcm based on your app’s requirements. This flexibility allows you to optimize for file size or quality without needing post-processing tools.
- mp3 for broad compatibility
- wav for high-quality archives
- opus for efficient storage
- flac for lossless quality
- pcm for raw processing
Language Support Beyond Hindi
The API supports 23 languages, allowing you to serve diverse dining audiences without managing separate voice models. European options include English, French, German, Spanish, and Italian, while Asian support covers Chinese, Japanese, Korean, and Hindi. Every voice can speak any of these languages, ensuring consistent audio quality across your platform’s international menu items.
English and European Languages
We support major European languages like English, French, German, Spanish, and Italian. You can use any of the eight voices to generate content in these languages. This flexibility helps you maintain a uniform brand voice across different regional menus.
- Supports 23 languages total
- Any voice works in any supported language
- No extra cost for multi-language content
Asian Languages Support
Our catalog includes Chinese, Japanese, Korean, Hindi, and other Asian languages. This allows you to localize menus for Asian dining platforms efficiently. We also support Arabic, Turkish, and Slavic languages for broader coverage.
Auto-Detection Limits
The API guesses the language from the script if you do not set the language field. Latin text defaults to English. For non-English Latin-script text, you must explicitly set the language parameter to avoid errors. Vietnamese, Indonesian, and Thai are not supported.
- Latin text defaults to English
- Set
languagefor non-English Latin text - Vietnamese, Indonesian, and Thai are missing
Pricing for Small Teams
Our pricing model scales directly with your usage, charging $0.01 per 1,000 characters. This transparent structure ensures you only pay for the audio you generate, making it ideal for small teams managing fluctuating menu updates or promotional announcements.
There are no monthly subscription fees, seat limits, or minimum spend requirements. You retain full control over your budget without the overhead of recurring billing cycles or complex enterprise contracts.
Pay Per Character
The rate is $10 per million characters, equivalent to $0.01 per 1,000 characters. Spaces and punctuation count toward this total, providing predictable costs for long-form menu descriptions or short snack alerts.
- $0.01 per 1,000 characters
- Spaces and punctuation count
- Failed requests are free
No Monthly Fees
We do not offer monthly plans or subscriptions. Your balance does not expire, and you can top up whenever needed without committing to a recurring cycle.
- No monthly plans
- No seats or minimums
- Balance never expires
Free Trial Credit
Sign in with Google to receive $0.50 in trial credit, valid for 7 days. Accounts created with email and password start without trial credit. This provides approximately 50,000 characters for testing voice options before committing to top-ups.
- $0.50 trial credit
- Valid for 7 days
- Google sign-in only

Voice Licensing and Rights
Svajpa Editorial provides synthetic voices rooted in licensed recordings, ensuring your dining platform’s audio features are built on a clear legal foundation without the overhead of custom voice design or expensive licensing negotiations.
Our catalog relies on established datasets, allowing indie developers to focus on integration rather than rights management for standard voice usage in restaurant menus and notifications.
LibriTTS-R Foundation
All eight voice presets are generated from the LibriTTS-R dataset, which carries a CC BY 4.0 license. This foundation ensures consistent quality across languages like Hindi, English, and German without requiring additional per-voice licensing for your API integration.
- Built on CC BY 4.0 licensed data
- Consistent engine quality across all 8 presets
- No extra licensing fees for standard usage
Public Domain Source
The underlying recordings originate from LibriVox public-domain audiobook readings. Because the source material is in the public domain, you do not need to track individual narrator rights for the base voice models used in your dining platform.
- Sourced from LibriVox public-domain readings
- No complex narrator rights tracking required
- Clean legal basis for synthetic generation
Inaudible Watermarking
Every generated audio file includes an inaudible watermark that marks the content as synthetic speech. This provides a transparent layer of identification for your users, distinguishing AI-generated menu descriptions or announcements from human-recorded audio.
- Inaudible watermark on all generated files
- Clear indication of synthetic speech
- No visible changes to audio quality
Performance and Reliability
Generating audio is straightforward: send a request and receive a file. Long texts are split into 4096-character parts and joined. Typical speed is roughly real-time, processing about 25 characters of text per second of audio. Short phrases may take a few seconds to synthesize.
Real-Time Generation
Requests are processed sequentially. Under normal load, generation is roughly real-time, but requests are processed sequentially and may wait in a queue under load. Each request receives its own audio file, which is returned directly in the response.
Queue Handling
Requests are processed one after another. If the system is under heavy load, your request may wait in a queue before synthesis begins. This ensures consistent quality but can increase latency during peak times.
Character Limits
Each request accepts up to 4096 characters. Spaces and punctuation count toward this limit. For longer menus or descriptions, split the text into parts and join the resulting audio files in your application.
Integration Workflow
Getting your dining platform’s audio features running requires minimal overhead. You begin by securing a unique API key from your dashboard, which serves as the sole credential for all requests. This key allows you to generate speech directly from your backend without managing complex authentication flows.
The workflow is designed for indie developers who prefer simplicity. You can test connectivity immediately using standard tools, ensuring your Hindi or English menu items render correctly before full integration.
API Key Setup
Navigate to /get-api-key/ to generate your first key. The interface displays your balance and character usage clearly. Store this key securely in your environment variables. If needed, you can issue a new key at any time, instantly invalidating the previous one.
- One active key per account
- Instant invalidation on renewal
- Visible balance and usage stats
cURL Examples
Test your configuration with a simple POST request. Send your JSON body to the speech endpoint to receive an audio file. This method works regardless of your programming language, making it ideal for quick validation of voice selection and format settings.
Use this approach to verify that your Hindi text is processed correctly before building your main application logic.
- POST to /v1/audio/speech
- JSON body format
- Immediate audio response
SDK Configuration
Leverage the OpenAI SDK by changing the base_url to https://api.svajpa.net/v1. This allows you to use familiar methods for generating speech. The request shape remains identical, so your existing code adapts with minimal changes.
This compatibility ensures you can integrate audio features without learning a new proprietary client library.
- Use OpenAI SDK
- Change base_url only
- Same request shape
Customer Support Model
Support for Svajpa Editorial’s API is handled exclusively via email, ensuring developers receive detailed, written responses without the noise of live chat interruptions. This model prioritizes clear, traceable communication for technical issues, billing inquiries, and key management requests.
We do not offer phone support, live chat widgets, or SLA-backed uptime guarantees. Our team focuses on asynchronous resolution, allowing you to continue building while we address your query with precision and context.
Email Only
All support tickets are managed through a single email channel. You will receive a structured response that includes relevant context, such as request IDs or character counts, ensuring no detail is lost in conversation threads.
- Written records for every inquiry
- Asynchronous resolution at your pace
- Context-rich responses with request IDs
No Live Chat
We have chosen to omit real-time chat widgets to reduce overhead and maintain focus on quality. This means you won’t encounter wait times in a queue or fragmented conversations, but you also won’t get instant answers during peak hours.
Direct Developer Contact
Your emails go directly to the development team, not a generic support desk. This ensures that technical questions about SDK compatibility, voice selection, or rate limits are answered by those who understand the API’s architecture.
Expect responses that reference specific endpoint behaviors or character limits, helping you debug integration issues without needing to explain the broader context of your dining platform.
- Direct access to technical team
- Context-aware troubleshooting
- No generic support scripts
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. The voice is neutral yet deep, avoiding excessive emotion or brightness.
23 languages
Any voice speaks any of these languages.
