A persistent personality-driven AI Telegram bot that embodies Rimuru Tempest from the Tensura universe.
Built with production-grade Node.js architecture by the Anitoku Team.
- Persistent Identity: Rimuru Tempest personality that never breaks character.
- Dynamic Relationships: Evolving user relationships from stranger to friend.
- Memory System: Remembers every conversation and user detail.
- Language Matching: Automatically responds in the user's language (Uzbek, English, Russian, etc.).
- Group Chat Support: Responds only when mentioned in group chats.
- API Key Rotation: Automatic Groq API key switching on rate limits (supports multiple keys).
- Atomic Data Storage: Corruption-safe local JSON database for chats, users, and relationships.
- Idle Follow-ups: Sends messages to friends after periods of silence.
- Node.js: v18.0.0 or higher.
- Telegram Bot Token: Created via @BotFather.
- Groq API Keys: At least one key from Groq Console.
# 1. Clone the repository
git clone https://github.com/TomsLiderUZ/rimuru-tempest-ai-telegram-bot
cd rimuru-tempest-ai-telegram-bot
# 2. Install dependencies
npm install
# 3. Create a .env file from the example
cp .env.example .env
# 4. Initialize data folders
npm run init-data
# 5. Start the bot
npm startCreate a .env file in the root directory and fill it with the following template:
# Telegram Configuration
BOT_TOKEN=your_bot_token
BOT_USERNAME=your_bot_username
ADMIN_NAME=Your_Admin_Name
ADMIN_CONTACT_LINK=https://t.me/your_admin_username
# AI Configuration (Groq API Keys)
GROQ_API_KEY_1=your_groq_key_1
GROQ_API_KEY_2=your_groq_key_2
GROQ_API_KEY_3=your_groq_key_3
GROQ_API_KEY_4=your_groq_key_4
GROQ_API_KEY_5=your_groq_key_5
# Model Configuration
GROQ_MODEL=openai/gpt-oss-120b
DEFAULT_LANGUAGE=uz
# Server Configuration
PORT=3000| Variable | Description |
|---|---|
BOT_TOKEN |
Your Telegram bot token. |
BOT_USERNAME |
The username of your bot (without @). |
ADMIN_NAME |
Your name for Rimuru to recognize. |
ADMIN_CONTACT_LINK |
Link to your Telegram for support. |
GROQ_API_KEY_1...N |
Multiple keys for rotation when rate limits are hit. |
GROQ_MODEL |
The AI model to use (default: openai/gpt-oss-120b). |
DEFAULT_LANGUAGE |
Default response language if detection fails. |
Rimuru's behavior changes based on his "opinion" of you:
- Stranger: Cold, dismissive, or suspicious. He will investigate how you reached him.
- Known: Recognizes you by name, becomes slightly more conversational.
- Friend: Warm, shares details about Tempest, sends follow-up messages if he hasn't heard from you.
- Enemy: Blocked status. Rimuru will refuse to talk to you after repeated insults.
The bot is programmed with "Mortal Pride" and "Dimensional Awareness". He knows he's connected via a "bridge" (portal) and will react with curiosity or condescension when challenged about his nature.
/start- Initial setup and introduction./restart- Manually restart the Telegram connection.
The bot includes a built-in health check server for platforms like Render, Heroku, or Railway.
- Port: The bot listens on the port defined by
$PORT(default 3000). - Endpoints:
GET /- Simple alive check.GET /health- JSON status with uptime and timestamp information.
- Persistency: If using a platform with an ephemeral filesystem, ensure your
data/directory is either on a persistent volume or use a cloud database provider.
# Repair the local JSON database if corrupted
npm run repair-db
# Run with development auto-restart
npm run devsrc/
├── ai/ # AI client, key rotation logic, dynamic prompt builder
├── bot/ # handlers, middleware, idle message scheduler
├── config/ # System settings, paths, and shared constants
├── memory/ # User storage, relationship states, conversation buffers
├── prompts/ # Rimuru's personality rules (merged core & behavior)
├── services/ # Higher-level logic orchestration
├── storage/ # Atomic JSON writer for safe data persistence
└── utils/ # Shared logger, time, and validation utilities
https://t.me/rimuru_tempest_ai_bot
MIT © ANITOKU Team