Casablanca – Anthropic has rolled out a new feature for its Claude Code tool that lets developers message their AI agent directly through apps like Telegram and Discord, shifting how people interact with coding assistants.
The update, called Channels, connects Claude Code to external messaging platforms so users can send instructions and receive replies while away from their desks. Instead of sitting in front of a terminal and waiting for responses, developers can now trigger tasks remotely and get notified when the work is done.
It builds on Anthropic’s Model Context Protocol, an open standard introduced in 2024. The system acts like a bridge between the AI and outside tools. When a message is sent through Telegram or Discord, it enters the running Claude session as an event. The agent can then write code, run tests, or fix bugs before sending a reply back through the same channel.
For this to work, the Claude session needs to stay active in the background, either in a terminal or on a server. Messages only arrive while the session is running. In practice, that turns Claude from a reactive chatbot into something closer to a persistent assistant.
Read also: Samsung Plans $73 Billion Investment Push in 2026 to Compete in AI Chip Race
The feature is still in research preview and requires Claude Code version 2.1.80 or later, along with the Bun runtime. It also only works with claude.ai accounts, not API keys, and must be enabled manually for team and enterprise users.
Setting it up involves creating a bot in Telegram or Discord, installing an official plugin inside Claude Code, and pairing the account using a generated code. Access is restricted through allowlists, meaning only approved users can send commands.
Anthropic also released a local testing tool called Fakechat, which runs a browser-based chat interface without needing external services. It’s meant to help developers understand how the system behaves before connecting real messaging platforms.
The move brings Claude closer to what open-source projects like OpenClaw have been offering for months. Those tools gained traction by letting users run always-on AI agents that could handle tasks like writing emails, building apps, or managing workflows from messaging apps.
But they often required complex setups and raised security concerns, especially when given deep access to personal systems.
By integrating similar capabilities into a managed product, Anthropic is betting on ease of use and tighter controls. The core model remains proprietary, but the connectors are built on open infrastructure, allowing developers to expand the system further.
Early reactions online suggest the gap between open-source agents and commercial AI tools is closing fast.
Morocco World News is also on X — check out our latest posts now! Get MWN on iOS and Android for instant access to breaking news.