Getting Started

Overview

Copy page

Build powerful speech applications with our Text-to-Speech and Speech Recognition APIs.

Services

Feature	Description
Text-to-Speech	Convert text to natural speech with Arabic and English voices
Speech Recognition	Transcribe audio to text with multi-language support
API Reference	Interactive API documentation with playground
Performance	Performance tuning and optimization guide
API Limits	Service limits and best practices

---

Getting Started in 5 Minutes

Get up and running with SM-AI-MODELS in just a few minutes.

Prerequisites

• SM-AI-MODELS services running locally or on a server
• curl or any HTTP client

1. Check Service Health

First, verify the services are running:

Code
 
# Check TTS service
curl http://localhost:9999/health

# Check ASR service
curl http://localhost:8088/health

Expected response:

Code
 
{"status": "healthy"}

2. Generate Your First Speech

Convert text to speech with a simple POST request:

Code
 
curl -X POST http://localhost:9999/v1/audio/speech \
  -H "Content-Type: application/json" \
  -d '{"input": "مرحباً بكم في يونيكود سولوشنز", "voice": "Yara"}' \
  --output welcome.mp3

This will create a welcome.mp3 file with the synthesized speech.

3. Transcribe Audio

Convert an audio file to text:

Code
 
curl -X POST http://localhost:8088/v1/audio/transcriptions \
  -F "file=@recording.wav"

Response:

Code
 
{"text": "مرحباً بكم في خدمات يونيكود سولوشنز"}

---

Service Endpoints

Both services provide REST and gRPC APIs for flexibility and performance.

Service	REST Port	gRPC Port	Description
SM-TTS-V1	9999	50051	Text-to-Speech service
SM-STT-V1	8088	50052	Speech Recognition service

API Options:

• REST API — Simple HTTP JSON endpoints, easy testing with cURL
• gRPC API — High performance, type-safe, streaming support

Learn more: gRPC Guide

---

Available Voices

Voice	Language	Description
Yara	Arabic	Female voice (default)
Nouf	Arabic	Female voice
Yara_en	English	Female voice

Try Different Voices

Arabic - Yara (Default)

Code
 
curl -X POST http://localhost:9999/v1/audio/speech \
  -H "Content-Type: application/json" \
  -d '{"input": "مرحباً", "voice": "Yara"}' \
  --output yara.mp3

Arabic - Nouf

Code
 
curl -X POST http://localhost:9999/v1/audio/speech \
  -H "Content-Type: application/json" \
  -d '{"input": "مرحباً", "voice": "Nouf"}' \
  --output nouf.mp3

English - Yara_en

Code
 
curl -X POST http://localhost:9999/v1/audio/speech \
  -H "Content-Type: application/json" \
  -d '{"input": "Hello, welcome!", "voice": "Yara_en"}' \
  --output english.mp3

Custom Audio Format

Request a different audio format:

Code
 
curl -X POST http://localhost:9999/v1/audio/speech \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Testing WAV format",
    "voice": "Yara_en",
    "response_format": "wav"
  }' \
  --output audio.wav

Supported formats: mp3, wav, opus, flac

---

Guides & Documentation

Guide	Description
REST API Reference	Complete TTS and ASR endpoint documentation
Streaming API	WebSocket and gRPC streaming for real-time audio
cURL Examples	Comprehensive cURL and grpcurl examples
gRPC Guide	Use gRPC APIs for high performance
Python Integration	Python code examples
Node.js Integration	JavaScript/TypeScript examples
Error Handling	Handle errors properly

---

gRPC Health Check

Both services implement the standard gRPC health check protocol:

Code
 
grpc.health.v1.Health/Check

Returns:
- SERVING     — Engine is ready
- NOT_SERVING — Engine failed to initialize

---

Made with ❤️ using UNICODE