RememberOrDismember Bot 🐟

A Telegram bot that helps you remember things 🗡️

Try it: @remember_or_dismember_bot

Features

• One-time Reminders: Set reminders for specific dates and times
• Recurring Reminders: Set up periodic reminders with cron-like scheduling
• AI-Powered Conversations: Powered by DeepSeek AI to infer cron expressions from natural language
• Job Management: Create, list, and cancel reminder jobs
• Webhook Support: Receives updates via webhooks for better performance
• Graceful Shutdown: Proper cleanup of resources and background jobs

Commands

• /start - Show help menu and bot introduction
• /newjob - Create a new reminder job (guided setup)
• /listjobs - List all your active reminder jobs
• /canceljob-<jobID> - Cancel a specific job (e.g., /canceljob-123)

Prerequisites

• Go 1.24+
• PostgreSQL database
• Telegram Bot Token (from @BotFather)
• DeepSeek API Key
• sqlc for database code generation

Install Dependencies

# Install sqlc for database code generation
brew install sqlc

# Install Go dependencies
go mod vendor && go mod tidy

Environment Setup

Create a .env file in the project root with the following variables:

ENV=dev
TELEGRAM_BOT_TOKEN=your_telegram_bot_token_here
DATABASE_URL=your_db_url_here
BASE_URL=https://your-domain.com
DEEP_SEEK_API_KEY=your_deepseek_api_key_here

Database Setup

The bot uses PostgreSQL with the following tables:

• chats: Stores chat information and context
• jobs: Stores reminder jobs with scheduling information

Database migrations are handled via SQL schema files in db/schemas/.

Local Development

1. Generate database code:

   sqlc generate

2. Set up your environment:

   • Ensure your PostgreSQL database is running and accessible

3. Run the bot:

   go run main.go

4. Set up webhook (for local development):

   • Use ngrok or similar tool to expose your local server
   • Update BASE_URL in your .env file

Build and Deploy

Google Cloud Run Deployment

1. Follow the basic setup here: https://cloud.google.com/run/docs/quickstarts/build-and-deploy/deploy-go-service

2. Amend cloudbuild.yaml with gcr.io/<PROJECT_ID>/<YOUR_SERVICE_NAME>

3. Run gcloud builds submit at the project root

4. Navigate to Cloud Run on the GC Console and deploy the submitted build

Docker

The project includes a Dockerfile for containerized deployment:

docker build -t remembertelebot .
docker run -p 9000:9000 --env-file .env remembertelebot

Architecture

• Bot Framework: Telegram Bot API with webhook support
• Database: PostgreSQL with sqlc for type-safe queries
• Job Scheduling: River queue for background job processing
• AI Integration: DeepSeek AI for conversational features
• Caching: Ristretto for in-memory caching
• Logging: Structured logging with zerolog

Project Structure

├── bot/                 # Telegram bot client
├── config/             # Configuration management
├── db/                 # Database schemas and generated code
│   ├── queries/        # SQL queries
│   ├── schemas/        # Database schema files
│   └── sqlc/          # Generated Go code
├── services/           # Business logic
│   ├── commands/       # Command handlers
│   ├── messages/       # Message handlers
│   └── callbackqueries/ # Callback query handlers
├── riverjobs/          # Background job processing
├── deepseekai/         # AI integration
├── ristrettocache/     # Caching layer
└── main.go            # Application entry point

License

This project is open source and available under the MIT License.