What audio quality and formats does Fonada V1 support?

HTTP /tts/generate-audio-large returns a complete WAV file (audio/wav, pcm_s16le, 24000 Hz, mono) ready to play. WebSocket /tts/generate-audio-ws streams raw PCM — initialize your audio stream with pcm_s16le, 24000 Hz, 1 channel. Languages: Hindi, Tamil, Telugu, and English.

Which languages are supported?

Fonada V1 supports 4 languages and 111+ system voices: Hindi 18, Tamil 15, Telugu 60, English 18. Code-mixing between Hindi and English is supported.

How fast is the text-to-speech processing?

Fonada V1 is optimized for low latency in real-time conversations and customer support. Processing speed depends on text length.

How do I integrate Fonada V1 into my application?

Use the REST API with your API key. A Python SDK is available. See the documentation for code examples.

What are the best practices for optimal speech quality?

Use proper punctuation, consistent spelling, avoid excessive special characters, keep requests under 450 characters, set the correct language, and test samples before large volumes.

What are the licensing terms and commercial usage rights?

Full commercial rights are included: complete ownership of generated audio, use in ads, podcasts, audiobooks, IVR, and apps, with no attribution, royalties, or usage limitations.

Fonada V1 TTS API Docs - System Voices Guide

Overview

Fonada V1 turns text into lifelike audio with natural intonation and clear pronunciation using pre-built catalog voices. Send a JSON request to the TTS endpoint with a voice name and a language label — no setup or training required. You can try it in the TTS playground by selecting the Fonada V1 model.

Generate natural media campaigns & ads
Stream real-time audio from text

Need voice cloning?

To synthesize speech with community / cloned voices from Voice Arena, see the Klone V2 Pro documentation.

Listen to a sample:

Sample Audio

Hindi Female Voice

00:15

"क्या आपने कभी सोचा है कि अगर हर ज़रूरी जानकारी, मौसम से लेकर ताज़ा खबरों तक, सिर्फ़ एक कमांड पर मिल जाए, तो आपका रोज़मर्रा का जीवन कितना आसान और मज़ेदार हो सकता है?"

Explore our voice library to find the perfect voice for your project.

Developer quickstart

Learn how to integrate text to speech into your applications

API Usage

Generate high-quality speech from text using pre-built system voices. Send JSON to the TTS endpoint with a voice name and language label.

Character Limit

Maximum 450 characters per request. For longer content, split your text into multiple API calls.

Generate Audio - Basic Request

curl

curl -X POST "https://api.fonada.ai/tts/generate-audio-large" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"input": "क्या आपने कभी सोचा है कि अगर हर ज़रूरी जानकारी, मौसम से लेकर ताज़ा खबरों तक, सिर्फ़ एक कमांड पर मिल जाए, तो आपका रोज़मर्रा का जीवन कितना आसान और मज़ेदार हो सकता है?", "voice": "Dhruv", "language": "Hindi"}' \
  --output output_large.wav

Request Parameters

Parameter	Type	Required	Description
input	string	✓	Text to convert to speech. Maximum 450 characters per request.
voice	string	✓	Voice name, e.g. Dhruv, Vaanee
language	string	✓	Hindi, English, Tamil, or Telugu

Response

You receive a complete, ready-to-play .wav file. The WAV header is already embedded — no manual stream initialization needed.

Content-Type: audio/wav
Encoding: pcm_s16le · Sample rate: 24000 Hz · Channels: 1 (mono)
Content-Length: [file_size]

Use --output filename.wav to save directly to file

API Endpoint

Base URL: https://api.fonada.ai
Endpoint: /tts/generate-audio-large
Method: POST
Content-Type: application/json

WebSocket Usage

Generate real-time streaming audio from text using our WebSocket endpoint for low-latency applications.

📡 WebSocket Endpoints

Production: wss://api.fonada.ai/tts/generate-audio-ws

WebSocket Request Parameters

Parameter	Type	Required	Description
api_key	string	✓	Your Fonadalabs API key for authentication
input	string	✓	The text you want to convert to speech
voice	string	✓	Voice name to use for synthesis (choose from 111+ Fonada V1 system voices across Hindi, Tamil, Telugu, and English)
language	string	✓	Language for speech synthesis ("Hindi", "Tamil", "Telugu", "English")

WebSocket Response Format

The WebSocket streams raw PCM audio in real-time chunks (not a WAV file):

Data Type: Raw PCM

Encoding: pcm_s16le

Sample Rate: 24000 Hz

Channels: 1 (mono)

Streaming: Audio data is sent progressively as it's generated

Binary bytes are pcm_s16le, 24000 Hz, 1 channel — initialize your audio stream with these params when using the WebSocket endpoint.

WebSocket Benefits

• Real-time streaming: Audio starts playing while text is still being processed
• Lower latency: Perfect for conversational AI and live applications
• Efficient bandwidth: Progressive audio delivery reduces waiting time
• Better UX: Users hear audio output immediately, improving perceived performance

Available Voices

Choose from our extensive collection of 111+ high-quality voices optimized for Hindi, Tamil, Telugu, and English languages.

Note: Use the voice name exactly as shown in your API requests, e.g., "voice": "Dhruv"

Hindi Voices (18)

Tamil Voices (15)

Telugu Voices (60)

English Voices (18)

Using Different Voices

For Hindi Content:

{"input": "नमस्ते", "voice": "Dhruv", "language": "Hindi"}

For Tamil Content:

{"input": "வணக்கம்", "voice": "Vaani", "language": "Tamil"}

For Telugu Content:

{"input": "నమస్కారం", "voice": "Naadamu", "language": "Telugu"}

For English Content:

{"input": "Hello, welcome", "voice": "Dhruv", "language": "English"}

Voice quality

Fonada V1 provides high-quality, natural-sounding speech synthesis with support for Hindi, Tamil, Telugu, and English languages, optimized for real-time conversational applications.

Fonada V1

Standard text-to-speech

System voices

Pre-built catalog voices with Hindi, Tamil, Telugu, and English — JSON REST API, 450 characters per request. Always returns one WAV file.

API endpoint: /tts/generate-audio-large
Languages: 4 (Hindi 18, Tamil 15, Telugu 60, English 18)
Voice param: voice name + language label

Supported formats

Both endpoints use the same PCM layout. How you consume it depends on the endpoint.

HTTP — `/tts/generate-audio-large`

You receive a complete .wav file. Headers are already embedded — no manual stream initialization needed.

Encoding: pcm_s16le

Sample rate: 24000 Hz

Channels: 1 (mono)

Content-Type: audio/wav

WebSocket only — `/tts/generate-audio-ws`

Binary bytes are pcm_s16le, 24000 Hz, 1 channel — initialize your audio stream with these params when using the WebSocket endpoint.

Encoding: pcm_s16le

Sample rate: 24000 Hz

Channels: 1 (mono)

Format: Raw PCM (no WAV header)

Supported languages

Currently 4 languages on Fonada V1:

Hindi, Tamil, Telugu, English

Note: We support code-mixing between English and Hindi, which is commonly used in Indian conversations.

Text Input

Fonada V1 converts plain text into natural-sounding speech. Simply provide your text input, and the model will generate high-quality audio output with natural intonation and pronunciation.

Current Capabilities

Natural speech synthesis from plain text
Support for English, Hindi, Tamil, and Telugu languages
High-quality audio output with clear pronunciation
Optimized for conversational applications

Coming Soon: Future versions will include advanced features like emotional context, voice modulation, and custom voice styles. Stay tuned for updates!

FAQ

HTTP: Complete WAV file (audio/wav) — pcm_s16le, 24000 Hz, 1 channel. Play or save as-is; no stream setup needed.
WebSocket: Raw PCM chunks — pcm_s16le, 24000 Hz, 1 channel. Initialize your audio stream with these params.
Quality: Natural prosody with crystal-clear pronunciation
Languages: Hindi, Tamil, Telugu, and English

Fonada V1

Overview

Need voice cloning?

Listen to a sample:

Developer quickstart

API Usage

Character Limit

Generate Audio - Basic Request

Request Parameters

Response

API Endpoint

WebSocket Usage

📡 WebSocket Endpoints

WebSocket Request Parameters

WebSocket Response Format

WebSocket Benefits

Available Voices

Hindi Voices (18)

Tamil Voices (15)

Telugu Voices (60)

English Voices (18)

Using Different Voices

Voice quality

Fonada V1

Supported formats

HTTP — `/tts/generate-audio-large`

WebSocket only — `/tts/generate-audio-ws`

Supported languages

Text Input

Current Capabilities

FAQ

What audio quality and formats does Fonada V1 support?

Which languages are supported?

How fast is the text-to-speech processing?

How do I integrate Fonada V1 into my application?

What are the best practices for optimal speech quality?

What are the licensing terms and commercial usage rights?

On this page

We use cookies

Fonada V1

Overview

Need voice cloning?

Listen to a sample:

Developer quickstart

API Usage

Character Limit

Generate Audio - Basic Request

Request Parameters

Response

API Endpoint

WebSocket Usage

📡 WebSocket Endpoints

WebSocket Request Parameters

WebSocket Response Format

WebSocket Benefits

Available Voices

Hindi Voices (18)

Tamil Voices (15)

Telugu Voices (60)

English Voices (18)

Using Different Voices

Voice quality

Fonada V1

Supported formats

HTTP — /tts/generate-audio-large

WebSocket only — /tts/generate-audio-ws

Supported languages

Text Input

Current Capabilities

FAQ

What audio quality and formats does Fonada V1 support?

Which languages are supported?

How fast is the text-to-speech processing?

How do I integrate Fonada V1 into my application?

What are the best practices for optimal speech quality?

What are the licensing terms and commercial usage rights?

On this page

We use cookies

HTTP — `/tts/generate-audio-large`

WebSocket only — `/tts/generate-audio-ws`