AsyncFakeYou

AsyncFakeYou

AsyncFakeYou is an asynchronous Python library for interacting with the FakeYou Text-to-Speech API. It allows you to submit TTS requests, track their status, and download the resulting audio files, all in an asynchronous manner using asyncio, aiohttp and aiofiles.

Features

• Asynchronous requests to FakeYou TTS API.
• Generation with parallel execution control
• Handle user authentication with session cookies

Installation

pip install asyncfakeyou

Simple usage

You can obtain a direct URL to the generated audio file.

from asyncfakeyou import AsyncAudioGen
import asyncio


async def text_to_speech():
    gen = AsyncAudioGen(cookies="your_cookie_string")
    audio_url = await gen.fetch_audio("model_token_here", "sample_text")
    print(audio_url)


asyncio.run(text_to_speech())

Or you can download the generated audio file to a specified directory.

from asyncfakeyou import AsyncAudioGen
import asyncio


async def text_to_speech():
    gen = AsyncAudioGen(cookies="your_cookie_string")
    audio_url = await gen.fetch_and_save_audio("model_token_here", "sample_text",
                                               output_path="./audio",
                                               filename="generated_audio.wav")
    print(audio_url)


asyncio.run(text_to_speech())

How to obtain cookies?

The cookies parameter of the AsyncAudioGen class is optional. However, I recommend setting it as it will give you higher queue priority, even without a paid subscription. If you have a premium account, you will be able to take advantage of all its benefits through this API.

from asyncfakeyou import receive_cookies
import asyncio


async def get_my_cookies():
    cookies = await receive_cookies("your_username_or_email", "your_password")
    print(cookies)


asyncio.run(get_my_cookies())

Parallel execution

If you need to generate multiple audio files, you can speed up the process by using fetch_tasks and fetch_and_save_tasks. These methods handle a fixed number of tasks in parallel (concurrent_tasks parameter) and automatically retry failed requests. The default value for concurrent_tasks is 3, but you can tweak this parameter.

You can iterate through the direct URLs to the generated audio files.

from asyncfakeyou import AsyncAudioGen
import asyncio


async def multiple_text_to_speech():
    gen = AsyncAudioGen(cookies="your_cookie_string")
    audio_tasks = [
        ("model_token_1", "sample_text_1"),
        ("model_token_2", "sample_text_2"),
        ("model_token_3", "sample_text_3")
    ]
    async for audio_url in gen.fetch_tasks(audio_tasks):
        print(audio_url)


asyncio.run(multiple_text_to_speech())

Or you can download the generated audio files to a specified directory.

from asyncfakeyou import AsyncAudioGen
import asyncio


async def multiple_text_to_speech():
    gen = AsyncAudioGen(cookies="your_cookie_string")
    audio_tasks = [
        ("model_token_1", "sample_text_1", "filename1.wav"),
        ("model_token_2", "sample_text_2", "filename2.wav"),
        ("model_token_3", "sample_text_3", "filename3.wav")
    ]
    await gen.fetch_and_save_tasks(audio_tasks, output_path="./audio")


asyncio.run(multiple_text_to_speech())

Contributing

Contributions are welcome! Please fork the repository and create a pull request with your improvements.