Environment Variables MilesONerd AI uses environment variables for configuration. All settings are stored in a .env file in the project root.
Required Variables .env.example
.env (example)
# Telegram Bot Token (get it from @BotFather) TELEGRAM_BOT_TOKEN = your_bot_token_here # API Keys (will be used in future implementations) SERPAPI_API_KEY = your_serpapi_key_here # Model Configuration (will be used in future implementations) DEFAULT_MODEL = mixtral-8x7b-v0.1 ENABLE_CONTINUOUS_LEARNING = true
Variable Descriptions TELEGRAM_BOT_TOKEN (Required) Your Telegram Bot API token. This is mandatory for the bot to function.
Open Telegram and search for @BotFather
@BotFather is Telegram’s official bot for creating and managing bots.
Create a new bot
Send the /newbot command to @BotFather and follow the prompts:
Choose a name for your bot (e.g., “MilesONerd AI Assistant”)
Choose a username ending in “bot” (e.g., “milesonerd_ai_bot”)
Copy your token
@BotFather will provide your bot token. It looks like: 1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
Copy this token to your .env file. Keep your bot token private! Anyone with this token can control your bot. Never commit it to version control or share it publicly.
SERPAPI_API_KEY (Optional) API key for SerpAPI Google Search integration. Currently planned for future implementation.
This feature is not yet implemented in the current version. You can leave this blank for now.
DEFAULT_MODEL (Optional) Specifies which AI model to use by default. Options:
llama - Llama 3.1-Nemotron (default)bart - BART model
The bot uses this setting in ai_handler.py:
self .default_model = os.getenv( 'DEFAULT_MODEL' , 'llama' )
Leave this as llama for best general-purpose performance. BART is automatically used for summarization tasks.
ENABLE_CONTINUOUS_LEARNING (Optional) Enables or disables continuous learning capabilities. Values: true or false (default: true).
self .enable_learning = os.getenv( 'ENABLE_CONTINUOUS_LEARNING' , 'true' ).lower() == 'true'
Continuous learning features are planned for future implementation.
Model Configuration The bot uses two AI models configured in ai_handler.py:
Llama 3.1-Nemotron Used for text generation and conversational responses:
'llama' : { 'name' : 'nvidia/Llama-3.1-Nemotron-70B-Instruct-HF' , 'type' : 'causal' , 'task' : 'text-generation' }
Capabilities:
General text generation
Conversational responses
Short and long-form answers
Context-aware replies
BART Used for text summarization:
'bart' : { 'name' : 'facebook/bart-large' , 'type' : 'conditional' , 'task' : 'summarization' }
Capabilities:
Text summarization
Long message condensation
TL;DR generation
Loading Configuration The bot loads environment variables using python-dotenv:
from dotenv import load_dotenv # Load environment variables load_dotenv()
This automatically reads the .env file when the bot starts.
Verifying Configuration The bot validates critical configuration on startup:
# Get the token from environment variable token = os.getenv( "TELEGRAM_BOT_TOKEN" ) if not token: logger.error( "No token found! Make sure to set TELEGRAM_BOT_TOKEN in .env file" ) return
If TELEGRAM_BOT_TOKEN is not set, the bot will exit with an error message. Ensure this variable is configured before starting the bot.
GPU Configuration The AI handler automatically detects GPU availability:
logger.info( f "CUDA available: { torch.cuda.is_available() } " ) if torch.cuda.is_available(): logger.info( f "GPU Device: { torch.cuda.get_device_name( 0 ) } " ) logger.info( f "Available GPU memory: { torch.cuda.get_device_properties( 0 ).total_memory / 1e9 :.2f} GB" )
Models automatically use GPU if available:
self .models[ 'bart' ] = BartForConditionalGeneration.from_pretrained( self .model_configs[ 'bart' ][ 'name' ], device_map = 'auto' if torch.cuda.is_available() else None , torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32, local_files_only = False )
No manual GPU configuration needed! The bot automatically uses GPU (float16) if available, otherwise falls back to CPU (float32).
Example Configuration File Here’s a complete example .env file:
# Bot credentials TELEGRAM_BOT_TOKEN = 1234567890:ABCdefGHIjklMNOpqrsTUVwxyz # Future features SERPAPI_API_KEY = # Model settings DEFAULT_MODEL = llama ENABLE_CONTINUOUS_LEARNING = true
Next Steps After configuration:
