Edge Tts

Installation

pip install edge-tts

Add to requirements.txt:

edge-tts>=7.0.0

Environment Config

# .env
EDGE_TTS_NAME=id-ID-ArdiNeural    # Voice name (see voice list below)

Quick Patterns

Generate Audio + SRT (Python — recommended)

import asyncio
import edge_tts
from pathlib import Path

async def generate_tts(
    text: str,
    output_audio: Path,
    output_srt: Path,
    voice: str = "id-ID-ArdiNeural",
    rate: str = "+0%",
    volume: str = "+0%",
    pitch: str = "+0Hz",
) -> None:
    """Generate MP3 audio and word-level SRT from text."""
    output_audio.parent.mkdir(parents=True, exist_ok=True)
    communicate = edge_tts.Communicate(
        text=text,
        voice=voice,
        rate=rate,
        volume=volume,
        pitch=pitch,
    )
    await communicate.save(str(output_audio), str(output_srt))

# Usage
asyncio.run(generate_tts(
    text="Tahukah kamu bahwa Titanic punya kolam renang?",
    output_audio=Path("output/scene_1/audio_1.mp3"),
    output_srt=Path("output/scene_1/subtitle_1.srt"),
))

Generate Audio Only (no SRT)

import asyncio
import edge_tts
from pathlib import Path

async def tts_audio_only(text: str, output: Path, voice: str = "id-ID-ArdiNeural") -> None:
    communicate = edge_tts.Communicate(text=text, voice=voice)
    await communicate.save(str(output))

asyncio.run(tts_audio_only("Hello world", Path("hello.mp3")))

CLI Usage

# Basic generation with subtitles
edge-tts --text "Hello, world!" --write-media hello.mp3 --write-subtitles hello.srt

# With voice selection
edge-tts --voice id-ID-ArdiNeural --text "Halo dunia!" --write-media hello_id.mp3

# Adjust rate, volume, pitch
edge-tts --rate=-20% --volume=+10% --pitch=-10Hz --text "Slower and deeper" --write-media adjusted.mp3

# List all available voices
edge-tts --list-voices

List Voices Programmatically

import asyncio
import edge_tts

async def list_voices(language_prefix: str = "id-ID") -> list[dict]:
    voices = await edge_tts.list_voices()
    return [v for v in voices if v["ShortName"].startswith(language_prefix)]

voices = asyncio.run(list_voices("id-ID"))
for v in voices:
    print(f"{v['ShortName']:30s} {v['Gender']:8s} {v.get('Locale', '')}")

Batch TTS for Multiple Scenes

import asyncio
import edge_tts
from pathlib import Path

async def batch_tts(
    scenes: list[dict],  # [{"text": "...", "audio": Path, "srt": Path}, ...]
    voice: str = "id-ID-ArdiNeural",
    rate: str = "+0%",
) -> None:
    for scene in scenes:
        print(f"Generating TTS: {scene['audio']}")
        comm = edge_tts.Communicate(text=scene["text"], voice=voice, rate=rate)
        await comm.save(str(scene["audio"]), str(scene["srt"]))

# Example
scenes = [
    {
        "text": "Scene satu narasi...",
        "audio": Path("output/scene_1/audio_1.mp3"),
        "srt": Path("output/scene_1/subtitle_1.srt"),
    },
    {
        "text": "Scene dua narasi...",
        "audio": Path("output/scene_2/audio_2.mp3"),
        "srt": Path("output/scene_2/subtitle_2.srt"),
    },
]
asyncio.run(batch_tts(scenes))

Popular Indonesian Voices

Popular English Voices

Rate / Volume / Pitch Reference

SRT Output Format

Edge TTS generates word-level SRT timing. Example output:

1