Frequently asked questions
What does a TTS API do?
A Text-to-Speech API converts text into audio files. You send a POST request with your text, and the API returns an audio file like MP3 or WAV. It is a simple cloud service: input text, receive audio. No infrastructure management is required, making it easy to add voice to any application.
How is it different from a plugin?
Unlike a plugin that installs directly into your app’s interface, a TTS API is a backend service. You call it via HTTP requests from your code. This decouples voice generation from your UI, allowing you to generate audio on your server and play it anywhere, such as in mobile apps or web clients.
Can I use it for chatbots?
Yes, VoiceGrid is ideal for chatbots. You can send user text to the API and receive audio to play back in your conversation flow. With support for 23 languages and multiple voices, it helps create engaging voice assistants without managing complex audio infrastructure.
Does it support real-time audio?
No, the API does not support real-time streaming or WebSocket connections. It returns the complete audio file after synthesis finishes. For long texts, you must split the input into chunks, generate each part, and join the resulting audio files in your application logic.
What formats does it output?
The API supports five audio formats: mp3, wav, opus, flac, and pcm. You specify the desired format in the request body. This flexibility allows you to choose the best balance between file size and quality for your specific chatbot or voice assistant implementation.
Is it expensive to use?
Pricing is straightforward: $10 per 1 million characters. Failed requests are free. You can start with a $0.50 free trial via Google sign-in. Top-ups are done via USDT on TRON, with no monthly plans or minimum spend, making it cost-effective for indie developers.
How many languages are supported?
VoiceGrid supports 23 languages, including English, Spanish, German, French, Japanese, and Arabic. Any voice can speak any supported language. If you do not specify a language, the system guesses it from the script, but setting it explicitly is recommended for non-English Latin text.
Can I customize the voice?
You cannot customize voice parameters like emotion or pitch beyond speed. You choose from 8 preset voices (e.g., Aria, Felix, Atlas). You can adjust the speed between 0.25 and 4.0. There is no custom voice cloning or uploading of your own voice recordings.
Is the audio synthetic?
Yes, the audio is synthetic. Every generated file contains an inaudible watermark marking it as synthetic speech. The voices are based on licensed recordings from LibriTTS-R and LibriVox, but the output is generated by the API engine, not recorded live for each request.
Does it use my data for training?
No, your input text is used only to produce the audio and is not used for training models. Generated audio is returned in the response and not stored. Only request metadata like time and character count is kept for billing purposes.
How do I integrate it?
Integration is simple: send a POST request to the API endpoint with your text and API key. The OpenAI SDKs work by changing the base URL. You can get your API key on the /get-api-key/ page after signing up with Google or email.
What is the character limit?
Each request supports up to 4096 characters. For longer texts, you must split the text into parts, send multiple requests, and join the resulting audio files. This ensures efficient processing and allows you to manage large scripts or long-form content effectively.
Can I change the speed?
Yes, you can adjust the speech speed between 0.25 and 4.0. This allows you to make the voice sound slower for clarity or faster for quick updates. The speed parameter is part of the request body and applies to the generated audio file.
How do I handle errors?
If synthesis fails, the charge is returned in full. You should check the API response for error codes and retry if necessary. Rate limits are 60 requests per minute per key. Monitor your usage on the dashboard to avoid hitting limits during high traffic.
How TTS APIs work
To use a Text-to-Speech API, you start by sending your text payload to our cloud endpoint. You structure a JSON body containing the input text, the desired model (like tts-1), and the target voice. This request is authenticated via your unique API key, ensuring secure access to the synthesis engine.
Send text to server
Your application sends a POST request to https://api.playh95.com/v1/audio/speech. The payload includes up to 4096 characters of text, the selected voice preset, and optional parameters like speed or language. This standard JSON format ensures compatibility with existing HTTP clients and OpenAI SDKs.
- Supports
tts-1andtts-1-hdmodels - Text limit of 4096 characters per request
- Standard JSON body structure
Server generates audio
Upon receiving your request, our servers process the text through our TTS engine. The system synthesizes the audio, applying the chosen voice characteristics and speed settings. If the input exceeds the character limit, the text is split into manageable parts for sequential processing.
Receive audio file
The server returns the generated audio file in your preferred format (mp3, wav, opus, flac, or pcm). You receive the file directly in the response, ready to be played or stored. Failed requests incur no charge, ensuring you only pay for successful synthesis.
Why use an API instead of a plugin
A plugin often locks you into a specific framework or version, forcing you to wait for updates or deal with compatibility conflicts. An API gives you direct control over every request, response, and error state, letting you build exactly the workflow your app needs without middleware overhead.
With a TTS API, you decide when and how audio is generated. You aren't tied to a plugin's internal logic or update cycle.
Full control over integration
You manage the request lifecycle directly. Send text via standard HTTP POST and handle the audio response exactly as your architecture requires, whether that's caching or immediate playback.
This eliminates the 'black box' problem of plugins, giving you transparent visibility into latency and errors for better debugging.
- Direct HTTP/JSON control
- Custom error handling
- No framework lock-in
Works with any language
Plugins are often built for specific languages or frameworks (e.g., React-only). An API works with any language that can send HTTP requests, from Python and Node.js to C++ and Rust.
You aren't limited by the plugin's ecosystem; if your bot is written in Go, it can use VoiceGrid just as easily as a JavaScript app.
Scalable architecture
As your chatbot grows, a plugin might bottleneck your server. An API separates concerns, allowing you to scale audio generation independently from your core application logic.
With 60 requests per minute per key and simple queuing, you can design your own scaling strategy without waiting for a plugin vendor to add features.

Common use cases
A Text-to-Speech API allows your application to convert written text into audible speech programmatically. Instead of pre-recording every possible phrase, your chatbot or voice assistant sends text to an endpoint and receives an audio file instantly. This enables dynamic, context-aware responses that feel personal and engaging.
This approach is essential for building voice interfaces that adapt to user input in real-time. By integrating the API, you gain full control over the audio generation process, ensuring your product scales efficiently without managing complex infrastructure.
Chatbot responses
Modern chatbots benefit significantly from TTS integration, moving beyond static text bubbles to deliver spoken conversations. Your bot can send user queries to the API and receive synthesized audio responses, creating a more immersive experience. With 8 distinct voice presets and support for 23 languages, you can tailor the tone to match your brand identity perfectly.
The API handles up to 4096 characters per request, allowing for detailed, nuanced responses without truncation. Failed requests are free, ensuring you only pay for successful audio generation, which keeps operational costs predictable even during high-traffic periods.
- Supports 23 languages with auto-detection or explicit setting
- 8 voice presets for varied tonal qualities
- Generates audio files instantly via REST API
Voice assistants
Voice assistants rely on low-latency speech synthesis to maintain natural conversation flow. By using the TTS API, developers can integrate high-quality synthetic voices into their assistants without building custom audio infrastructure. The API’s straightforward JSON interface works with standard HTTP clients, making integration straightforward for indie developers.
Speed settings between 0.25x and 4.0x allow you to fine-tune the delivery rate for clarity or energy. Whether you need a calm instructional voice or an energetic promo style, the 8 presets cover a wide range of use cases, from news updates to casual chit-chat.
- Adjustable speed from 0.25x to 4.0x
- Supports real-time interaction patterns
- Compatible with standard HTTP clients
Accessibility features
TTS APIs are crucial for making digital content accessible to users with visual impairments or reading difficulties. By converting text content into audio, you ensure that your application complies with accessibility standards without requiring manual audio recording for every piece of content.
The API returns audio in multiple formats, including MP3 and WAV, ensuring compatibility with various screen readers and assistive technologies. Since input text is not used for training, user data remains private, providing a secure option for sensitive content delivery.
- Multiple output formats for broad compatibility
- Privacy-focused: input text not used for training
- Supports screen readers and assistive tech
Real-time vs batch processing
Text-to-Speech APIs bridge the gap between raw text and audible speech by processing requests through a cloud endpoint. Instead of managing local servers, you send JSON payloads and receive audio files, enabling seamless integration into chatbots, voice assistants, and accessibility tools without infrastructure overhead.
Real-time for chatbots
For chatbot applications, immediate response is critical. The API generates audio files; for long texts, generation takes roughly real-time, but short phrases may take a few seconds. This instant feedback loop keeps interactions natural and engaging, turning static text replies into dynamic voice conversations.
- Audio file generation for conversational flow
- Audio file delivery suitable for live user interaction
- Seamless integration with chatbot frameworks
Batch for large projects
Batch processing is ideal for pre-generating audio for e-learning courses, audiobooks, or extensive documentation. You can send multiple requests to process large volumes of text offline. This approach decouples audio creation from user interaction, ensuring consistent quality without impacting real-time performance.
- Pre-generate audio for offline playback
- Ideal for long-form content like audiobooks
- Offloads processing from live user sessions
Queue handling
Requests are processed sequentially, meaning higher loads may introduce slight delays. Under heavy traffic, a request might wait in a queue before synthesis begins. Despite this, typical speeds remain efficient, roughly matching real-time generation rates, ensuring predictable delivery even during peak usage.
- Sequential processing model
- Potential queue delays under high load
- Predictable real-time generation speed
Voice customization options
VoiceGrid gives you immediate control over how your synthetic audio sounds, letting you fine-tune the delivery without complex engineering. You select from eight distinct presets, adjust the playback rate, and define the language to match your exact use case.
Select preset voices
Choose from eight licensed voices, ranging from warm narrators like Aria to deep documentary tones like Orion. Each voice is based on real recordings, offering distinct personalities for your chatbots or assistants without needing custom cloning.
- Aria: Warm, clear narrator
- Felix: Energetic, promo
- Atlas: Neutral, news
Adjust speed
Set the speed parameter between 0.25 and 4.0 to control pacing. Slower speeds improve clarity for accessibility, while higher speeds compress information for quick updates in voice assistants.
- Range: 0.25–4.0
Choose language
Explicitly set the language field using an ISO code for precise control, especially with Latin-script text that might otherwise default to English. Supports 23 languages, with any voice capable of speaking any supported language.
- 23 languages supported
- Auto-detection available
Audio format considerations
To choose the right response_format, balance file size against fidelity. MP3 is the standard for web delivery, offering small file sizes that load quickly in chat interfaces without sacrificing intelligibility for most conversational use cases.
MP3 for small size
MP3 provides efficient compression suitable for mobile data usage and rapid playback in voice assistants. It remains the default choice for most indie integrations where bandwidth matters.
WAV for high quality
WAV delivers uncompressed audio, preserving every nuance of the synthetic voice. This format is ideal for archival purposes or when downstream processing requires raw audio data without decoding artifacts.
OPUS for efficiency
OPUS offers superior compression efficiency compared to MP3, maintaining clarity at lower bitrates. It is excellent for low-bandwidth environments where data conservation is critical.

Integration complexity
Integrating VoiceGrid requires only a standard HTTP POST request to our API endpoint. You send your text payload in JSON format, and the server processes it synchronously to return an audio file. This simplicity means you don’t need complex middleware or specialized libraries to get started immediately.
The architecture supports both lightweight direct calls and structured SDK integrations. Since the request shape mirrors the OpenAI speech endpoint, you can easily adapt existing codebases. Just update your base URL and inject your API key to begin generating speech within minutes.
Simple REST calls
Sending a POST request to https://api.playh95.com/v1/audio/speech is all you need. Provide the model, input text, voice, and format in a JSON body. The server returns the audio file directly in the response, making it easy to save to your application’s storage.
- Standard HTTP POST method
- JSON request body
- Direct audio file response
SDK support
We don’t maintain a proprietary SDK, but our API is fully compatible with the official OpenAI SDKs. By simply changing the base_url to https://api.playh95.com/v1, you can use your existing Python, JavaScript, or Go code without rewriting the core logic. This reduces integration time significantly for developers already familiar with common TTS patterns.
- Compatible with OpenAI SDKs
- No custom library needed
- Easy base URL swap
Documentation availability
Our documentation at /docs/ provides complete reference materials, including endpoint details, parameter descriptions, and code examples in multiple languages. You can also test the API directly via the /get-api-key/ page, which includes a ready-to-use cURL example. This ensures you have all the technical details needed for a smooth integration.
- Complete API reference
- Code examples included
- Interactive testing available
Cost implications
VoiceGrid charges strictly by usage: $10 per 1 million characters, which equals just $0.01 per 1,000 characters. Spaces and punctuation count toward this total, so you pay only for what you generate. Failed requests are free, ensuring you never pay for errors.
There are no monthly plans, seats, or minimum spend requirements. You simply top up your balance and spend it as needed. This model eliminates wasted budget on unused capacity or unused seats in your team.
Pay per character
The cost is transparent: $10 per 1 million characters. This rate applies to all models, including tts-1 and tts-1-hd.
Every character in your input text, including spaces and punctuation, contributes to the total count. You can track exact usage via your dashboard.
- $0.01 per 1,000 characters
- Spaces and punctuation count
No upfront costs
Start with a $0.50 free trial (about 50,000 characters) valid for 7 days when signing in with Google.
No credit card is required to start, and there are no subscriptions or recurring fees to maintain your account.
Predictable billing
Balance does not expire, and you can top up with USDT (TRC-20) on the TRON network. Minimum top-up is $5.
Request metadata is kept for billing, but your input text is not used for training, ensuring clear separation between usage and data.
Quality of synthetic speech
VoiceGrid delivers synthetic speech that sounds natural and engaging, leveraging licensed recordings from LibriTTS-R to provide clear, human-like audio for your applications.
Natural sounding voices
Our 8 preset voices, such as Aria’s warm narration or Felix’s energetic promo style, are based on licensed recordings, ensuring high-quality output without the need for complex configuration.
Distinct personalities
While we do not offer emotion tags or custom voice cloning, each voice has a distinct character. You can select Sage for calm instructional content or Orion for deep documentary styles, giving your project a consistent tonal identity.
Clear articulation
- Input text is not used for training
- Generated audio is not stored
Language support scope
VoiceGrid supports 23 languages, including English, Spanish, Mandarin, Arabic, and Swahili. Every voice can speak any of these languages, allowing you to build truly global applications without managing separate voice models for each region.
We provide robust auto-detection and explicit setting options to ensure accurate pronunciation and natural flow, regardless of your user base's linguistic diversity.
23 languages supported
Support includes Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, and Turkish. If a language is not in this list, it is not supported.
- 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 feature
When you do not specify a language, the system guesses it from the text script. Latin script text defaults to English.
Explicit language setting
For non-English Latin-script text, always set the language parameter to ensure correct pronunciation. This prevents the default English fallback from misinterpreting foreign words.

Data privacy and usage
Your input text is used strictly to generate the audio and is not used for training our models. This ensures your proprietary content remains yours. Generated audio is returned in the response and is not stored, providing clear data boundaries for indie developers integrating TTS into their apps.
We retain request metadata—such as time, character count, and status—solely for accurate billing. This minimal data keeps your usage transparent without retaining the actual voice content you generate.
No training on input
Input text is used only to produce the audio and is not used for training.
Generated audio is returned in the response and not stored.
Temporary storage
Request metadata (time, characters, status) is kept for billing.
This ensures accurate charges without retaining your voice content.
User control
You retain full control over your input data.
Billing records are transparent and based on usage only.
Listen to the 8 voices
Each sample below was generated by this API from the same sentence. Design conversational experiences that feel natural for users interacting with voice assistants and chatbots daily.
23 languages
Any voice speaks any of these languages.
