Kawn Docs

Overview

Sawt Najd (صوت نجد) is a state-of-the-art neural Text-to-Speech (TTS) model designed specifically for Arabic speech synthesis. It converts Arabic text into clear, expressive, and natural-sounding audio in WAV format.

Sawt Najd is ideal for:

Conversational AI & Voice Agents — Power interactive voice response (IVR) and AI assistants with natural Arabic voices.
Content & Article Narration — Automatically generate spoken audio from articles, books, and Islamic educational materials.
Accessibility — Provide read-aloud functionality for visually impaired users.
E-Learning & Pronunciation — Generate accurate Arabic speech for language learning and educational platforms.

Endpoint

POST /v1/tts

Synthesizes speech from the provided Arabic text prompt and returns a binary WAV audio stream (audio/wav).

Requires an API Key passed via the x-api-key header or Authorization: Bearer <YOUR_API_KEY>.

Request Body

Field	Type	Required	Description
`model`	`string`	Yes	The TTS model to use. See Models.
`prompt`	`string`	Yes	The Arabic text content to convert into speech.
`voice`	`string`	Yes	The voice identifier to use. See Voices.

Example Request

curl -X POST http://api.kawn.ai/v1/tts \
  -H "x-api-key: <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sawt-najd/voxcpm2",
    "prompt": "مرحباً بكم في منصة كَون، بوابتكم لنماذج الذكاء الاصطناعي للغة العربية والمحتوى الإسلامي.",
    "voice": "rashid"
  }' \
  --output speech.wav

Models

Model ID	Provider	Description
`sawt-najd/voxcpm2`	Sawt Najd	High-fidelity Arabic speech synthesis model with natural intonation and accurate pronunciation.

Voices

Voice ID	Gender	Description
`rashid`	Male	Warm and natural male voice, well suited for formal narration and conversational agents.
`fatimah`	Female	Clear and expressive female voice, ideal for assistant and narration use.
`default`	—	Standard default voice profile.

Response

The endpoint streams raw binary audio data with the following header:

Content-Type: audio/wav

You can stream the response directly to an audio player or save it to a .wav file.

Code Examples

curl -X POST http://api.kawn.ai/v1/tts \
  -H "x-api-key: <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sawt-najd/voxcpm2",
    "prompt": "السلام عليكم ورحمة الله وبركاته",
    "voice": "fatimah"
  }' \
  --output output.wav

import fs from 'node:fs';

const response = await fetch('http://api.kawn.ai/v1/tts', {
  method: 'POST',
  headers: {
    'x-api-key': process.env.KAWN_API_KEY,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    model: 'sawt-najd/voxcpm2',
    prompt: 'السلام عليكم ورحمة الله وبركاته',
    voice: 'fatimah',
  }),
});

if (!response.ok) {
  throw new Error(`HTTP error! status: ${response.status}`);
}

const arrayBuffer = await response.arrayBuffer();
fs.writeFileSync('output.wav', Buffer.from(arrayBuffer));
console.log('Audio saved to output.wav');

import os
import requests

url = "http://api.kawn.ai/v1/tts"
headers = {
    "x-api-key": os.environ.get("KAWN_API_KEY"),
    "Content-Type": "application/json",
}
payload = {
    "model": "sawt-najd/voxcpm2",
    "prompt": "السلام عليكم ورحمة الله وبركاته",
    "voice": "fatimah",
}

response = requests.post(url, json=payload, headers=headers)
response.raise_for_status()

with open("output.wav", "wb") as f:
    f.write(response.content)

print("Audio saved to output.wav")

Sawt Najd TTS