Frequently asked questions
What is an AI voice emulator?
Our service is a cloud text-to-speech API, not an emulator. It converts text into audio files using licensed recordings from LibriTTS-R. Each generated file includes an inaudible watermark to identify it as synthetic speech. We do not offer custom voice cloning or emotion tags, focusing instead on reliable, natural-sounding narration for your content.
How does AI text-to-speech work for news?
You send text via a POST request to our API endpoint. The system synthesizes the audio and returns the file. For longer articles, split the text into parts under 4096 characters and join the resulting audio files. This workflow allows indie developers to automatically turn news articles into natural audio without complex infrastructure.
Which AI voice is best for news reading?
Atlas is designed as a neutral, news-style voice, making it a natural fit for standard reporting. However, 'best' depends on your specific editorial tone. For calm, instructional segments, Sage works well, while Felix suits energetic promos. All eight voices support all 23 languages, so you can choose the voice that best matches your audience’s preference without technical constraints.
Can I change text to audio automatically?
Yes, via our REST API. Send a JSON body with your text, voice selection, and desired format to the /v1/audio/speech endpoint. You can automate this process in your news platform’s backend. The API returns the audio file directly, enabling seamless integration into your existing digital news workflows.
Is AI powered text to speech expensive?
Pricing is $10 per 1 million characters, or $0.01 per 1,000 characters. Failed requests are free. There are no monthly plans, seats, or minimum spends. You only pay for what you use, making it cost-effective for indie developers scaling their news audio output without predictable fixed costs.
How many characters can I send per request?
You can send up to 4096 characters per request. For longer articles, split the text into chunks of this size. The API processes each chunk and returns individual audio files, which you can then join. This ensures reliable synthesis even for lengthy news broadcasts or detailed reports.
What languages does the API support?
We support 23 languages including English, Spanish, French, German, Japanese, and Hindi. Any voice can speak any supported language. If you do not set the language parameter, the system guesses it from the script, defaulting to English for Latin text. Vietnamese and Thai are not supported.
Do I need to pay for failed requests?
No. If synthesis fails, the charge is returned in full. You only pay for successful audio generations. This ensures you are not penalized for temporary issues or invalid inputs. Failed requests do not consume your character balance, protecting your budget during high-volume news updates.
How fast is the audio generation?
Typical speed is roughly real time, generating about 25 characters per second. Short phrases may take a few seconds. Requests are processed sequentially, so under heavy load, a request may wait in a queue. This predictable timing helps you plan your news delivery schedules accurately.
Can I use the OpenAI SDK with this API?
Yes. Our API uses the same request shape as the OpenAI speech endpoint. You can use the official OpenAI SDK by changing the base_url to https://api.kurisumala.com/v1 and using your Curisum API key. This allows easy integration if your team is already familiar with OpenAI’s tooling.
What audio formats are available?
Choose from mp3, wav, opus, flac, or pcm. MP3 and Opus are efficient for web delivery, while WAV and FLAC offer high quality for archival. PCM is available for raw audio processing. Note that we return complete audio files, not streaming audio. Select the format that best suits your news platform’s requirements.
Is there a free trial available?
Yes. Signing in with Google grants $0.50 credit (approx. 50,000 characters), valid for 7 days. Email-based accounts start without trial credit. There are no subscriptions or minimum spends, allowing you to test the API risk-free before committing.
How do I top up my account?
Top up using USDT on the TRON network (TRC-20). The minimum amount is $5, and credits are automatically added within minutes after the transfer. We do not accept cards, PayPal, or invoices. Your balance does not expire, so you can fund your account whenever convenient for your news operations.
Do you store my input text?
We do not store your input text. It is used solely to generate audio and is not used for training. The generated audio is returned in the response and not stored. We retain request metadata—such as time, character count, and status—strictly for billing purposes. Your content remains private and transient during processing.
First request
curl
curl https://api.kurisumala.com/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.mp3How It Works
Curisum Audio transforms your article text into natural-sounding audio with a single API call. You send a JSON payload containing the text and voice preference, and we return the synthesized audio file ready for playback in your news application.
Send text, receive audio
The process is straightforward: submit your text via a standard HTTP request. The API processes the input and returns the audio file in your chosen format, allowing you to integrate speech synthesis directly into your content pipeline without managing complex infrastructure.
Simple REST endpoint
We provide a single, clean endpoint at https://api.kurisumala.com/v1/audio/speech using the POST method. You simply provide your authentication key, the text input, and desired parameters. This simplicity reduces integration time, letting indie developers focus on building their news platforms rather than debugging audio pipelines.
OpenAI SDK compatible
Our API follows the same request shape as the OpenAI speech endpoint. You can use the official OpenAI SDKs by simply changing the base_url to https://api.kurisumala.com/v1 and using your Curisum API key. This compatibility means you likely already have the code structure needed to get started quickly.
Voices for News Content
Curisum Audio provides eight distinct voice presets designed to suit various editorial needs. Each voice is derived from licensed recordings, ensuring a professional foundation for your news platform without the complexity of managing audio infrastructure.
Select voices that align with your content style, from neutral reporting to deeper documentary narration, ensuring consistent audio quality across all articles and updates.
Neutral news delivery
Atlas offers a neutral voice optimized for news content. This preset is ideal for delivering daily updates and factual reporting with clarity and consistency.
The voice maintains a steady pace suitable for standard news broadcasts, ensuring listeners receive information without unnecessary stylistic distraction.
- Ideal for standard news updates
- Consistent delivery for factual reporting
Documentary style depth
Orion provides a deep voice suited for documentary-style content. This preset adds weight to longer pieces, investigative reports, or historical recaps.
Its authoritative tone helps engage listeners with more detailed or narrative-driven journalism.
Use Orion when your content requires a serious, immersive listening experience that goes beyond standard reporting.
- Deep, documentary-style tone
- Suitable for investigative pieces
Friendly conversational tone
Milo delivers a friendly and conversational voice. This preset works well for community-focused news or lighter segments that benefit from a more personal touch.
It helps indie developers create a closer connection with their audience by making the audio feel approachable and engaging.
Use Milo for segments that require a warmer, less formal delivery style.
- Friendly and conversational
- Good for community news

Global Language Support
Curisum Audio powers news audio in 23 languages, from Arabic and Swahili to Japanese and Turkish. Our engine handles diverse scripts and dialects, ensuring your readers hear clear, natural speech regardless of their location.
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 can speak any of these languages, giving you flexibility in content production.
Automatic language detection
The API guesses the language from the text script if you do not specify one. Latin script defaults to English, but for non-English Latin text, we recommend setting the language explicitly to avoid misinterpretation.
Specific language codes
Pass an ISO language code in the request to force a specific voice output. This ensures consistent delivery for regional news variants, such as distinguishing between European and Brazilian Portuguese, even though our voices are universal presets.
Developer-Friendly Limits
The API accepts up to 4096 characters per request. For longer articles, split the text into chunks and join the resulting audio files. This simple structure keeps each payload manageable and predictable.
4096 character chunks
Each request processes a maximum of 4096 characters. Spaces and punctuation count toward this limit. If your news segment exceeds this, split the input text and concatenate the output audio files.
Rate limits explained
You are allowed 60 requests per minute per API key. If your balance is zero, the limit is 10 requests per minute. Requests are processed sequentially, so heavy load may cause brief queuing.
Real-time generation speed
Generation speed is roughly one second of audio per 25 characters of text for long passages. Short phrases may take a few seconds. Processing happens one request at a time.
Transparent Pricing
Pricing is straightforward: $10 per 1,000,000 characters, or $0.01 per 1,000 characters. Spaces and punctuation count toward the total. If a synthesis request fails, you are not charged, and any attempted charge is returned in full.
$10 per million characters
Each request processes up to 4,096 characters. For longer articles, split the text into chunks and join the resulting audio files. This keeps requests efficient and predictable.
No monthly subscriptions
There are no plans, seats, or minimum spends. Use USDT on the TRON network (TRC-20) to top up your balance. Minimum top-up is $5. Funds are credited automatically within minutes and never expire.
- No credit cards or PayPal
Free failed requests
If the API returns an error during synthesis, the request is free. You only pay for successfully generated audio files. This protects your budget during development or unexpected outages.

Technical Specifications
Curisum Audio supports multiple output formats to suit different deployment needs. You can choose MP3 for compact file sizes, WAV or FLAC for lossless quality, or PCM for raw data processing.
Multiple audio formats
Select MP3, WAV, OPUS, FLAC, or PCM via the response_format parameter. Each format serves specific use cases, from web delivery to archival storage.
Speed control options
Adjust playback speed between 0.25 and 4.0 using the speed field. This allows you to optimize listening duration without re-generating the audio file.
Watermarked synthetic audio
Every generated audio file contains an inaudible watermark identifying it as synthetic speech. This helps distinguish AI-generated news content from human-recorded broadcasts.
Why Indie Developers Choose Us
Curisum Audio offers a straightforward REST API that mimics the OpenAI request shape, allowing you to integrate TTS into your news platform without adopting a proprietary ecosystem. You can use the official OpenAI SDKs by simply pointing them to our base URL, ensuring a clean, standard integration path for indie developers.
No vendor lock-in
Our API follows the standard OpenAI request shape, so you can switch providers or add redundancy with minimal code changes.
Simple billing via crypto
We support standard authentication, making it easy to secure your endpoints within your news app.
- Top-up via USDT (TRC-20) with no monthly subscriptions or invoices.
- Balance does not expire, and failed requests are free.
Direct API access
Access our REST endpoints directly without waiting for third-party wrappers or proprietary SDKs. Use our official documentation to integrate quickly, whether you prefer curl, Python, or Node.js. Get your API key instantly upon sign-up and start sending text to audio requests immediately.
Integration Examples
Integration is straightforward because our endpoint mirrors the OpenAI speech structure. You can use existing OpenAI SDKs by simply pointing them to our base URL, allowing indie developers to swap in Curisum Audio without rewriting their audio generation logic.
Curl command example
Send a simple POST request with your API key. The endpoint accepts JSON bodies containing the model, input text, voice, and format preferences. You receive the audio file directly in the response body.
- POST https://api.kurisumala.com/v1/audio/speech
- Header: Authorization: Bearer tts-your-api-key
Python SDK integration
Use the official OpenAI Python library. Configure the base_url to https://api.kurisumala.com/v1 and inject your Curisum key. The audio.speech() method handles the request and returns the audio bytes or file path.
- Install:
pip install openai - Set
base_urlto our API endpoint
Node.js implementation
The OpenAI Node.js SDK works identically. Initialize the client with our base URL and key. Call the speech endpoint to generate audio files for your news articles, handling the response as a standard file download or buffer.
- Install:
npm install openai - Configure client with our API domain

Data Privacy & Usage
Curisum Audio uses your input text strictly for the immediate production of the audio file. We do not retain your articles to train or improve our speech models, ensuring your editorial content remains yours.
Text not used for training
Unlike many consumer TTS services, we do not use your submitted text to fine-tune our neural networks. Your content is processed solely to generate the requested audio file and is not stored for future model training.
No audio storage
The generated audio is returned directly in the API response and is not stored on our servers. Once you receive the file, it is gone from our infrastructure, giving you full ownership and control over the output.
Billing metadata retention
We retain minimal metadata—such as request timestamps, character counts, and status codes—strictly for accurate billing and usage tracking. This data is used only to verify that you are charged correctly for the characters generated.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. news outlets often overlook the creative spirit of indie developers in their coverage.
23 languages
Any voice speaks any of these languages.
