Kevlarbot AI

Multi-provider AI chatbot for Telegram

---

Features

• 6 free AI models (MiMo, Llama, Gemma, Qwen, Mistral, GPT-OSS)
• BYOK support for OpenRouter, DeepSeek, and Claude
• 5 personas: Default, Coder, Translator, Pirate, Tutor
• Session memory with persistent chat history
• Encrypted API keys (Fernet)
• Rate limiting (2s cooldown, 100 msgs/day)
• Admin tools: broadcast & stats

Quick Start

git clone https://github.com/KevClint/Kevlarbot-AI.git
cd Kevlarbot-AI
python -m venv venv
.\venv\Scripts\Activate.ps1
pip install -r requirements.txt
copy config.example.env config.env
python bot.py

Docker

git clone https://github.com/KevClint/Kevlarbot-AI.git
cd Kevlarbot-AI
docker compose up -d --build

New? Follow the full Setup Guide for detailed installation, configuration, and troubleshooting instructions.

Environment Variables

Variable	Required	Description
TELEGRAM_TOKEN	Yes	Bot token from @BotFather
ADMIN_IDS	No	Comma-separated Telegram user IDs
GROQ_API_KEY	No	For Llama, GPT-OSS models
HF_API_KEY	No	For Gemma, Qwen, Mistral models
MIMO_API_KEY	No	For MiMo V2.5 model

Commands

Command	Description
/start	Welcome message
/new	Clear memory, keep keys & model
/model	Switch AI engine
/persona	Change assistant behavior
/setkey <provider> <key>	Save BYOK API key
/info	View dashboard
/session	View current session
/retry	Regenerate last response
/cancel	Cancel in-progress request
/help	Help menu
/stats	View user count (admin)
/broadcast <msg>	Send to all users (admin)

Free Models (in telegram bot only)

Key	Model	Provider
gemma-2b	Gemma 2 2B	HuggingFace
qwen-3b	Qwen2.5 3B	HuggingFace
llama-8b	Llama 3.1 8B	Groq
mistral-nemo	Mistral Nemo 12B	HuggingFace
gpt-oss	GPT-OSS 20B	Groq
mimo	MiMo V2.5	MiMo

Personas

Key	Name	Behavior
default	Default Assistant	General-purpose helpful AI
coder	Expert Coder	Senior engineer, precise code
translator	English Translator	Translate to clear English
pirate	Sarcastic Pirate	Correct answers, pirate character
tutor	Academic Tutor	Step-by-step explanations

Project Structure

├── bot.py                  # Entry point
├── Dockerfile              # Docker build file
├── src/
│   └── kevlarbot/
│       ├── __init__.py
│       ├── handlers.py     # Command & callback handlers
│       ├── ai_client.py    # AI provider client
│       ├── database.py     # SQLite database layer
│       ├── providers.py    # Provider & persona definitions
│       ├── config.py       # Configuration & encryption
│       └── utils.py        # Utility functions
├── docs/
│   ├── README.md
│   └── setup.md
├── config.env              # Environment variables (git-ignored)
├── kevlarbot.db             # SQLite database (git-ignored)
├── requirements.txt
└── LICENSE

License

MIT