yottoCode

Introduction to yottoCode

yottoCode is a native macOS application that connects Claude Code with Telegram, built on the official Anthropic Agent SDK. It allows users to control their Claude Code agent directly from Telegram, providing a seamless and powerful interface for coding, project management, and AI-assisted tasks. The app offers full agent controls, including voice input and output, permission keyboards, session resume, and model switching—all integrated into the Telegram environment.

The product is designed for developers and power users who want to leverage the capabilities of Claude Code while maintaining the convenience and accessibility of Telegram. With support for multiple models (Haiku, Sonnet, Opus), MCP servers, and real-time streaming, yottoCode bridges the gap between desktop development and mobile interaction. It also provides advanced security features such as Telegram user ID whitelisting and inline permission controls, making it a secure and flexible solution for AI-powered workflows.

Takeaways

• Native macOS app powered by the official Anthropic Agent SDK
• Full control over Claude Code via Telegram
• Voice input and output with local processing
• Interactive permission keyboards and session resume
• Model switching between Haiku, Sonnet, and Opus
• MCP server integration and live cost tracking
• Secure access with Telegram user ID whitelist
• Supports unlimited bots and messages with Pro plan
• No API keys required; uses existing Claude Code subscription

How yottoCode Works

yottoCode operates by connecting to the Claude Code agent through the Anthropic Agent SDK. Users create a Telegram bot, configure the app with their bot token, and select a project directory. Once set up, messages sent to the Telegram bot are processed by the Claude Code agent, which can then perform actions like code generation, file manipulation, and task automation. All interactions are handled securely, with permissions managed through inline keyboards and session data persisted locally on the user's machine.

The app supports real-time streaming of responses, allowing users to see the AI's thought process as it generates code or performs tasks. It also includes features like voice input and output, where voice notes are transcribed using Whisper and responses are generated as voice replies—all processed locally on the user's Mac. This ensures that sensitive data remains on the user's device and is not transmitted to external servers.

Core Benefits and Applications