tulgatts 0.1.3

Uses Character AI for synthesis.

TulgaTTS

TulgaTTS - A library for synthesizing text-to-speech (TTS) with various popular voices in Kazakhstan.

Installation

# Via pip (in the future)

pip install tulgatts

# For audio playback (optional):

pip install -e .[audio]

Getting Started

1. Obtain your token:

   • ⚠️ DO NOT SHARE YOUR TOKEN! It is required to send requests from your account.

   • Open link https://character.ai/chat/2WPyJRflV_4nTx6_-tuNrhkiiqhDyOsn9O25BR1sDO8

   • Open Developer Tools in your browser (usually F12).

   • Go to the Network tab.

   • Refresh the page or send a message to any character.

   • In the list of network requests, click on any request (for example, to /settings).

   

   • Go to the Headers section.

   • Find the Authorization header. It will look like:

     
     Authorization: Token <your_token_here>
     
     

   • Copy only the token part (after Token ).

   • Set the CHARACTER_AI_TOKEN environment variable or create a .env file in the project root and add CHARACTER_AI_TOKEN=your_copied_token.

2. Use the library.

Usage Examples

You can use the library in several ways. For simple scripts, use the synchronous method; for async applications, use the async/await method.

1. The simplest usage

from tulgatts import TulgaTTS



# Create a client (default voice – Tokaev)

client = TulgaTTS(api_token="CHARACTER_AI_TOKEN", voice="Tokaev")



# 1. Speak with the default voice and save to file

client.say("This is an example created with this class")

2. Synchronous usage

import os

from tulgatts import TulgaTTS

from dotenv import load_dotenv



# Load token from .env

load_dotenv()

api_token = os.getenv("CHARACTER_AI_TOKEN")



def sync_say_usage():

    if not api_token:

        print("❌ No token!")

        return



    try:

        tts_client = TulgaTTS(api_token=api_token, voice='Tokaev')

        print(f"✅ Client ready with voice '{tts_client.voice}'.")



        tts_client.say("Synchronous example.", output_file="sync_example.mp3")

        print("🎧 File ready: sync_example.mp3")



    except Exception as e:

        print(f"⚠️ Error occurred: {e}")



if __name__ == "__main__":

    print("2. Synchronous example...")

    sync_say_usage()

    print("-" * 20)

3. Asynchronous usage

import os, asyncio

from tulgatts import TulgaTTS

from dotenv import load_dotenv



# Load token from .env

load_dotenv()

api_token = os.getenv("CHARACTER_AI_TOKEN")



async def advanced_async_say():

    if not api_token: return print("No token!")

    

    try:

        tts_client = TulgaTTS(api_token=api_token)

        print(f"Client ready with voice '{tts_client.voice}'.")

        await tts_client.say_async("Third example, part one.", voice="Tokaev", output_file="advanced_async1.mp3", play_audio=True)

    except Exception as e: print(f"Error: {e}")



if __name__ == "__main__":

    print("\n3. More advanced async example...")

    asyncio.run(advanced_async_say())

    print("All examples completed.")

Available Voices

The library comes with the following pre-configured voices:

• Nursultan Nazarbaev

• Tokaev

• Pavel Durov

• Nurlan Saburov

• Putin

• Skriptonit

• Elon Musk

• Albert E

• Yandex Alica

• J.A.R.V.I.S

• Tony Stark

Note: Sometimes TulgaTTS may not synthesize your expected text. This issue is being worked on.

Special thanks to PyCharacterAI for enabling TTS with Character AI voices.

Authors

• David Suragan (TulgaTTS)

• Gemini AI

License

MIT