IdeaBot

Telegram bot that captures text, voice notes, media, and forwards into an Obsidian vault. Optionally it can:

• transcribe audio via OpenAI-compatible STT API (for example local faster-whisper server)
• summarize notes via Ollama
• create tasks in Vikunja

Features

• Accepts Telegram messages, captions, forwards, voice and attachments
• Saves markdown notes to Obsidian with frontmatter and tags
• Stores media files (images/files) in configurable vault folders
• Optional daily note inbox insertion
• Optional deduplication state file
• Optional task sync to Vikunja

Project Structure

• bot.py - main Telegram bot logic
• requirements.txt - Python dependencies
• Dockerfile - image for the bot service
• docker-compose.yml - example stack (bot + optional STT + optional Ollama)
• .env.example - safe environment template (no real secrets)

Quick Start

1. Create bot via BotFather and get token.
2. Copy env template and fill real values:

cp .env.example .env

3. Create local data folders:

mkdir -p data/vault state data/hf-cache data/ollama

4. Start services:

docker compose up -d --build

5. Open Telegram and send /chatid to your bot, then set ALLOWED_CHAT_ID in .env.

Minimal Required Variables

• TELEGRAM_BOT_TOKEN
• ALLOWED_CHAT_ID
• OBSIDIAN_INBOX_DIR
• OBSIDIAN_DAILY_DIR

License

MIT