Cloud TTS API for Chatbots and Voice Assistants

By VoiceGrid editorial team · Updated 2026-11-18

VoiceGrid is a cloud Text-to-Speech API designed specifically for indie developers building chatbots and voice assistants. If you need high-quality synthetic speech without managing complex infrastructure, this is your solution. We provide the reliable audio engine that powers your voice applications, letting you focus on your product logic while we handle the synthesis.

On this page, you will find everything needed to integrate voice into your app quickly. We offer a simple REST API compatible with OpenAI SDKs, eight distinct voice presets, and support for 23 languages. Whether you are generating short responses or long monologues, get clear, consistent audio with transparent pricing and no hidden fees. Let’s build your voice assistant together.

Cloud TTS API for Chatbots and Voice Assistants

Frequently asked questions

What is VoiceGrid?

VoiceGrid is a cloud text-to-speech API designed for indie developers and chatbot creators. It allows you to send text and receive an audio file via a simple endpoint. We provide the reliable audio engine that powers your voice assistants without the infrastructure headache, enabling easy integration into your apps.

How does the TTS API work?

Send a POST request to the /v1/audio/speech endpoint with your text, voice, and format preferences. You receive an audio file in return. For longer texts, split input into 4096-character parts and join the resulting audio files. The API handles the synthesis, returning an audio file after a short delay.

Which voices are available?

We offer eight presets: Aria, Nova, Luna, and Sage (female); Felix, Milo, Atlas, and Orion (male). Voices are based on licensed recordings from LibriTTS-R and LibriVox. Each voice has a distinct character, from warm narrators to deep documentary styles, suitable for various assistant personalities.

How many languages are supported?

We support 23 languages, including English, Spanish, German, French, Japanese, and Chinese. Any voice can speak any supported language. If the language is not set, it is guessed from the script, defaulting to English for Latin text. Vietnamese, Thai, and Ukrainian are not supported.

What audio formats can I use?

You can choose from mp3, wav, opus, flac, and pcm formats. Specify the desired format in the request body. These options allow you to balance file size and quality according to your application needs, whether for web playback or local processing.

How fast is the generation?

Generation is roughly real-time on long texts, processing about 25 characters per second. Short phrases may take a few seconds. Requests are processed sequentially, so under heavy load, a request might wait in a queue. Speed depends on text length and current server load.

Can I use the OpenAI SDK?

Yes. Our endpoint shape matches the OpenAI speech endpoint. You can use the official OpenAI SDK by changing the base_url to https://api.playh95.com/v1 and using your VoiceGrid API key. This allows for easy integration if you are already familiar with OpenAI's client libraries.

Is there a free trial?

Yes, signing in with Google gives you $0.50 of credit, valid for 7 days. This covers about 50,000 characters. Accounts created with email and password start without trial credit. You can top up your balance anytime to continue using the service after the trial.

How do I pay?

We accept USDT on the TRON network (TRC-20) only. The minimum top-up is $5, and credits are added automatically within minutes. There are no monthly plans, no cards, and no subscriptions. Your balance does not expire, and you only pay for what you use.

What happens if a request fails?

If synthesis fails, the charge is returned in full. You are not billed for failed requests. This ensures you only pay for successfully generated audio files. Check your dashboard to see your usage and balance adjustments after any failed attempts.

Do you store my input text?

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. We keep request metadata, such as time, characters, and status, strictly for billing purposes.

Are the voices synthetic?

Yes, every generated file carries an inaudible watermark marking it as synthetic speech. The voices are based on licensed recordings from LibriTTS-R and LibriVox public-domain audiobook readings. They are not imitations of real people or celebrities, but distinct synthetic presets.

What is the rate limit?

The limit is 60 requests per minute per key. Before your first top-up, the limit is 10 requests per minute. If you exceed the limit, requests may be delayed. You can issue a new key at any time; the old one stops at once.

How do I get support?

Support is available by email only. There is no phone number, live chat, or SLA. We do not offer enterprise contracts or on-premise installations. For technical issues or billing questions, send an email to our support address, and we will assist you.

First request

curl

curl https://api.playh95.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.mp3

Why developers choose VoiceGrid

Sound wave illustration for: Why developers choose VoiceGrid

VoiceGrid provides a straightforward REST API that fits directly into your existing chatbot or voice assistant architecture. You simply send text and receive an audio file, making integration as easy as updating a single base URL in the official OpenAI SDKs.

Unlike building your own TTS engine, we handle the heavy lifting. You don’t need to manage servers, scale instances, or maintain audio infrastructure. This reliability means your voice assistants stay responsive and consistent without the operational headache.

Simple REST API

Our API follows a familiar request shape, allowing you to use standard libraries with minimal code changes. Send a POST request to the speech endpoint, and receive your audio file after a short delay (short phrases take a few seconds; longer texts take roughly real-time).

  • Standard JSON body with model, input, and voice parameters
  • Compatible with official OpenAI SDKs by changing base_url
  • Supports multiple audio formats including mp3, wav, and opus

No infrastructure to manage

Focus on your product logic, not server maintenance. VoiceGrid processes requests sequentially, so under load a request may wait in a queue. Whether you have ten users or ten thousand, the engine handles the load.

  • Automatic scaling for varying traffic loads
  • No server configuration or deployment needed
  • Consistent performance regardless of your app's size

Consistent audio quality

Every request is processed by the same high-quality engine, ensuring uniform tone and clarity across all your voice interactions. From short prompts to long narratives, the audio output remains stable and professional.

  • Uniform voice characteristics across all generations
  • High-fidelity audio suitable for voice assistants
  • Predictable output for reliable user experience

Built for chatbots and voice apps

Sound wave illustration for: Built for chatbots and voice apps

VoiceGrid is engineered for the demanding environment of voice assistants and interactive chatbots, where every millisecond of delay can break the illusion of conversation. Our API delivers audio files; generation takes a few seconds for short phrases and roughly real-time for long texts.

We understand that natural dialogue isn't always brief. Whether you are reading a detailed instruction or a multi-turn story, our system handles the nuances of voice synthesis with precision, ensuring that the transition between sentences feels natural and uninterrupted for your end-users.

Low latency responses

For short phrases, generation takes just a few seconds, providing immediate feedback that keeps the user experience snappy. For longer texts, the system processes audio at roughly real-time speed (about 25 characters per second of generation), ensuring that the audio output arrives at a pace that matches natural speech patterns.

Requests are processed sequentially, so under load, a request may wait in a queue. Requests are processed sequentially, so under load a request may wait in a queue.

Handle long conversations

Since a single request supports up to 4,096 characters, you can feed substantial blocks of dialogue without immediate truncation. For texts exceeding this limit, the recommended workflow is to split the input into parts, generate the audio for each segment, and join the files seamlessly.

This approach allows your chatbot to deliver rich, continuous audio experiences without hitting API character limits. By managing the chunking logic on your end or using standard integration patterns, you maintain full control over the narrative flow and audio continuity.

Support for multiple languages

VoiceGrid supports 23 languages, including English, Spanish, French, German, Japanese, Korean, Arabic, Hindi, and Russian. Any voice can speak any of these languages, giving you the flexibility to deploy your chatbot globally without managing separate voice models for each region.

If the language field is not set, the system guesses the language from the script (defaulting to English for Latin text). For non-English Latin-script text, it is best practice to explicitly set the language parameter to ensure accurate pronunciation and intonation from the start.

  • Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Russian, Spanish, Swahili, Swedish, Turkish
Built for chatbots and voice apps

8 distinct voice presets

Sound wave illustration for: 8 distinct voice presets

VoiceGrid provides eight distinct voice presets, each derived from licensed recordings and public-domain audiobook readings. These voices are engineered to fit specific conversational contexts, allowing indie developers to select the perfect tone for their chatbots or voice assistants without needing custom cloning or complex configuration.

Aria: warm narrator

Aria delivers a warm, clear narrative style ideal for explanatory content or guided interactions. Her tone remains consistent and professional, making her suitable for onboarding flows or informational chatbots where clarity is paramount. She handles the 4096-character limit per request; longer texts should be split into parts and joined.

  • Warm, clear narrator tone
  • Ideal for onboarding and explanations
  • Consistent performance over long texts

Nova: upbeat assistant

Nova brings a bright, energetic quality to your application. She is perfect for customer-facing bots that need to maintain a friendly, engaging demeanor. Whether handling quick queries or longer conversations, Nova’s preset ensures the interaction feels lively and responsive, fitting the modern expectation of helpful digital assistants.

  • Bright, upbeat energy
  • Great for customer-facing bots
  • Maintains engagement in quick queries

Luna: storytelling voice

Luna offers a soft, immersive tone designed for storytelling or relaxing audio experiences. Her preset is optimized for narratives, providing a soothing auditory backdrop that keeps users engaged. This makes her an excellent choice for apps focused on content consumption, such as news summaries or bedtime stories.

  • Soft, storytelling tone
  • Optimized for narratives
  • Soothing for content consumption

Support for 23 languages

Sound wave illustration for: Support for 23 languages

VoiceGrid supports 23 languages, ensuring your chatbot sounds natural whether addressing English speakers or global audiences. Any voice can speak any of these languages, giving you full creative control over your assistant's personality without needing multiple voice models.

English, Spanish, French

Start with widely spoken languages to reach a broad audience quickly. The system automatically detects Latin-script text as English, but you can explicitly set the language parameter for other Latin-based languages like Spanish or French to ensure perfect accent and pronunciation.

German, Japanese, Korean

Expand into complex scripts and tonal languages with precision. Whether you are building a voice assistant for Berlin, Tokyo, or Seoul, the engine handles the specific phonetic requirements of German, Japanese, and Korean seamlessly.

Arabic, Hindi, Russian

Reach diverse markets with robust support for Arabic, Hindi, and Russian. These languages are fully supported across all eight voice presets, allowing you to maintain consistent audio quality regardless of the target region.

Developer-friendly format

Sound wave illustration for: Developer-friendly format

VoiceGrid is designed to drop into your existing workflow with minimal friction. By mirroring the standard speech endpoint structure, you can leverage familiar tools and libraries without learning a new protocol from scratch.

OpenAI SDK compatible

Use the official OpenAI SDKs by simply pointing your base_url to https://api.playh95.com/v1. This allows you to keep your existing code structure while accessing our TTS engine with your own API key.

Multiple audio formats

Choose the output format that best fits your application architecture. We support MP3, WAV, OPUS, FLAC, and PCM, ensuring compatibility with everything from lightweight web players to high-fidelity voice assistants.

Easy integration

Sending a POST request with your text and voice preference returns an audio file directly. Combine this with our character-splitting logic to handle long dialogues seamlessly in your chatbot pipelines.

Developer-friendly format

Transparent pricing

Sound wave illustration for: Transparent pricing

VoiceGrid charges strictly by character count at a flat rate of $10 per million characters, meaning you pay only for what you use. Failed requests are free, so you never waste credits on errors. Spaces and punctuation count toward the total, ensuring predictable billing without surprise overages.

Pay per character

At $0.01 per 1,000 characters, pricing scales linearly with your usage. Whether you are generating short bot responses or long audiobooks, the cost remains consistent and easy to calculate. You can estimate your monthly spend by multiplying your expected character volume by this rate.

No hidden fees

There are no monthly plans, no per-seat charges, and no minimum spend requirements. You also do not need to worry about VAT invoices or subscription renewals. Payments are made via USDT (TRC-20) only; there are no cards, PayPal, or subscriptions.

Free trial included

Sign up with Google to receive $0.50 in free credit, which covers approximately 50,000 characters. This trial is valid for 7 days, giving you enough time to test the API quality. Email-based accounts start without trial credit but can top up immediately.

Reliable performance

Sound wave illustration for: Reliable performance

VoiceGrid is engineered to deliver audio quickly, even under load. Requests are processed sequentially, meaning longer texts may queue briefly, but short phrases return in seconds. For extended content, split text into 4096-character chunks to maintain smooth playback without overwhelming the engine.

Real-time generation

Synthesis typically runs at roughly real-time speed, generating about 25 characters of text per second of audio. While requests are queued one after another, short prompts return almost instantly, making it ideal for interactive chatbot responses where users expect immediate audio feedback.

Consistent output

Every request uses the same high-quality engine, ensuring your voice assistant sounds identical across all interactions. Whether you are generating a quick greeting or a long narrative, the audio quality remains stable, with no degradation based on queue depth or time of day.

Scalable API

The API handles 60 requests per minute per key, sufficient for most indie apps. If you exceed this limit, simply split larger texts into smaller segments. The system has a rate limit of 60 requests per minute per key (10 before the first top-up).

Data privacy focused

Sound wave illustration for: Data privacy focused

VoiceGrid is built with developer privacy in mind. We treat your input data with strict boundaries, ensuring that the text you send for synthesis is used solely for the immediate generation of audio. There is no secondary use of your conversational logs or prompts for model training or improvement.

No training on input

The text you submit via the API is processed exclusively to produce the requested audio file. It is not ingested into our training datasets, meaning your unique dialogue flows and content remain yours.

No audio storage

Generated audio is returned directly in the API response and is not persisted on our servers. Once the file is delivered to your application, the raw audio data is not stored, reducing your storage footprint and data exposure.

Clean data policy

We retain only the necessary metadata—such as request timestamps, character counts, and status codes—for accurate billing and debugging purposes. No unnecessary personal or content data is kept after the transaction is complete.

Data privacy focused

Get started in minutes

Sound wave illustration for: Get started in minutes

Start by signing up on the /get-api-key/ page using your Google account. This gives you immediate access to $0.50 in free trial credit, valid for 7 days, letting you test the API without entering payment details.

Sign up with Google

Visit the registration page and click the Google sign-in button. Your account is created instantly, and you receive trial credit to explore the voices.

Get your API key

Once signed in, navigate to your dashboard to view your unique API key. This key authenticates your requests to the TTS endpoint. You can issue a new key at any time, which invalidates the old one immediately.

  • Copy your key from the dashboard
  • Keep it secure for your app
  • Revoke anytime for fresh security

Make your first request

Use the provided cURL example or the OpenAI SDK with your base URL to send a POST request. Input your text, select a voice like 'nova', and receive the audio file directly. Failed requests are free, so you can test confidently.

  • Set base_url to https://api.playh95.com/v1
  • Send text via POST /v1/audio/speech
  • Receive MP3, WAV, or OPUS audio

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.

Avatar of the voice aria
ariafemale
Avatar of the voice nova
novafemale
Avatar of the voice luna
lunafemale
Avatar of the voice sage
sagefemale
Avatar of the voice felix
felixmale
Avatar of the voice milo
milomale
Avatar of the voice atlas
atlasmale
Avatar of the voice orion
orionmale

23 languages

Any voice speaks any of these languages.

Sources and further reading

  1. Wikipedia
  2. LibriVox public-domain audiobooks
  3. Python documentation
  4. curl documentation
  5. W3C WAI: making audio and video accessible
  6. openai-python on GitHub
  7. MDN: Web Speech API
  8. Creative Commons Attribution 4.0
  9. LibriTTS-R paper (arXiv 2305.18802)
  10. TRON developer hub: TRC-20
  11. LibriTTS-R corpus (OpenSLR 141)
  12. OpenAI API reference: create speech

More on this site