Telegram AI

An advanced plugin that syncs Telegram messages to your vault with AI-powered processing (OpenAI), smart categorization, and automated content organization.

✨ Key Features

• 🤖 AI Intelligence: Supports OpenAI (GPT-4 & Whisper). Automatically analyzes text, images, and seamlessly transcribes voice messages, audio, and video files.
• 📂 Local Processing: Extracts text from documents (PDF, DOCX, TXT, Code files) locally to protect privacy and save API costs, with AI processing as a fallback.
• 📸 Media Albums: Smartly handles Telegram media groups/albums, keeping context together in a single note.
• 🔗 Smart Logic: URL-only messages skip AI to save tokens; distinct prompts can be applied based on the content type.
• 🛡️ Robust Error Handling: Automatically detects and displays user-friendly alerts for API issues, such as exhausted billing quotas or invalid/revoked keys, preventing silent failures.
• 📝 Dynamic Templates: Use powerful variables like {{ai:title}}, {{category}}, and {{date:YYYY-MM}} for file naming.

🚀 Quick Start

1. Install: Download main.js, manifest.json, and styles.css from Releases to .obsidian/plugins/telegram-ai/.
2. Obsidian: Settings → Community Plugins → Enable "Telegram AI Sync".
3. Telegram Bot: Create a bot via @BotFather, copy the Token.
4. Configure:
   • Enter Bot Token in plugin settings.
   • Add your Telegram User ID to "Allowed Chats".
   • Select your AI Provider and enter the API Key.

📋 Configuration & Usage

Template Variables

Customize how notes are created in the settings:

• {{date:YYYY-MM-DD}}: Current date.
• {{ai:title}}: AI-generated title based on content.
• {{content}}: The processed message body.
• {{category}}: AI-detected category (e.g., Work, Ideas).

AI Prompts

Using the built-in wide-modal prompt editor, you can define specific behavior and toggle processing for different content types:

• General Formatting: Applied to all processed content as a final formatting step.
• Text: Instructions for processing plain text messages.
• Audio & Video: A unified prompt for handling transcripts generated by the Whisper API from voice messages, audio files, and video tracks.
• Photos: "Analyze this image and extract text/context..." (Requires Vision API).
• Documents: "Summarize this document..." (Handles parsed text from PDFs, DOCX, etc.).

🎛️ Settings Overview

Telegram Connection

• Bot Token: Your Telegram bot token.
• Allowed Chats: Whitelist of authorized users.
• Connection Status: Real-time connection monitoring.

AI Configuration

• Provider Selection: OpenAI.
• Model Settings: Temperature, max tokens, timeout.
• Prompt Management: Content-specific, unified media, and general prompts with a dedicated full-width editing interface.
• Processing Toggles: Enable/disable AI parsing for each content type (Text, Photos, Audio/Video, Documents).

Organization

• Categories: Define note categories with keywords.
• Templates: Customize file paths and naming.
• Distribution Rules: Advanced message routing.
• Local Processing: Enable direct document text extraction for PDF and DOCX files.

📢 Community

Join our Telegram channel for updates, tips, and support.

🤝 Acknowledgments

Special thanks to the open-source community.

• Foundation: This plugin was inspired by and built upon the excellent work of obsidian-telegram-sync by soberhacker.
• Libraries: Built with Obsidian API, node-telegram-bot-api, pdf-parse, and mammoth.