Text-to-Speech Generator Actor

An Apify actor that converts text into high-quality, natural-sounding speech using AI. It supports multiple languages and voices, is optimized for performance, and outputs professional-grade WAV audio files.

Overview

This actor is a text-to-speech (TTS) generator powered by advanced AI models. You provide text, and it returns a downloadable audio file. It's built for reliability and speed, with pre-loaded models and GPU acceleration support. Common use cases include creating voiceovers, audiobook narration, podcast content, accessibility tools, and multilingual audio applications.

Key Features

• Multi-language & Voices: Synthesizes speech in 9 languages (English, Spanish, French, Hindi, Italian, Portuguese, Japanese, Chinese) using over 50 distinct premium voices.
• Performance Optimized: Starts up quickly (~2 seconds) using pre-loaded models and automatically uses CUDA for GPU acceleration when available.
• Flexible Controls: Adjust speech speed from 0.1x to 3.0x the normal rate.
• Robust Processing: Handles long texts efficiently via automatic chunking and includes voice validation to match voices with the correct language.
• Professional Output: Generates broadcast-quality, uncompressed 24kHz WAV audio files.
• Detailed Metrics: Returns audio duration, file size, and processing performance data.

How to Use

Configure the actor's input via the Apify platform or API. The core required parameter is the text you want to convert.

Basic Input Configuration

Example Inputs

Basic English TTS:

{
  "text": "Welcome to our advanced text-to-speech system.",
  "voice": "af_heart",
  "lang": "en-us",
  "speed": "1.0"
}

Multilingual Synthesis:

{
  "text": "Bonjour et bienvenue.",
  "voice": "ff_siwis",
  "lang": "fr-fr",
  "speed": "1.2"
}

Voice Selection

Select a voice parameter from the lists below. The actor validates that the voice matches the specified language.

• English (US/GB): af_heart (F), am_adam (M), bf_alice (F), bm_daniel (M), and many more.
• Spanish: ef_dora (F), em_alex (M), pf_dora (F).
• French: ff_siwis (F).
• Hindi: hf_alpha (F), hm_omega (M).
• Italian: if_sara (F), im_nicola (M).
• Portuguese: pf_dora (F), pm_alex (M).
• Japanese: jf_alpha (F), jm_kumo (M).
• Chinese: zf_xiaoxiao (F), zm_yunxi (M).

Input/Output

Input

Provide input as a JSON object containing the parameters defined in the How to Use section.

Output

The actor's dataset contains a detailed JSON object for each run. The primary output includes:

• Audio File: A high-quality WAV format audio file available for download.
• Public URL: A direct, publicly accessible URL to the generated audio file.
• Audio Metadata: Duration (seconds), file size (bytes), and sample rate (24000 Hz).
• Configuration Used: The voice, language, and speed settings applied.
• Performance Metrics: Processing time and pipeline initialization time.

A typical output item structure:

{
  "audioUrl": "https://api.apify.com/v2/actor-runs/RUN_ID/datasets/DATASET_ID/items/ITEM_ID?attachment=audio.wav",
  "duration": 4.75,
  "fileSize": 571308,
  "format": "wav",
  "sampleRate": 24000,
  "voice": "af_heart",
  "language": "en-us",
  "speed": "1.0",
  "processingTime": 1.24
}