Anthropic simply shipped an OpenClaw killer known as Claude Code Channels, letting you message it over Telegram and Discord

The hit open supply autonomous AI agent OpenClaw could have simply gotten mogged by Anthropic.

As we speak, Anthropic introduced Claude Code Channels, a approach to hook up its personal highly effective Claude Code AI agentic harness to a human consumer's Discord or Telegram messaging functions, letting them message Claude Code immediately at any time when they need whereas on the go and instruct it to jot down code for them. Official documentation is right here.

This isn't only a new UI; it’s a basic shift in how builders work together with AI brokers, shifting from a synchronous "ask-and-wait" mannequin to an asynchronous, autonomous partnership. Beforehand, Claude Code customers had been caught interacting with the agentic harness on the Claude desktop software, terminal or supported developer setting, and Claude cellular app via a considerably flaky (in my expertise) interconnection setting known as Distant Management.

Now, Anthropic is providing among the similar core performance as OpenClaw that drove its speedy adoption amongst software program builders and vibe coders following its launch in November 2025 by Austrian developer Peter Steinberger (who, paradoxically, initially known as his challenge "Clawd" in honor of Anthropic's personal AI mannequin Claude which powered it initially, till Anthropic despatched him a cease-and-desist for potential trademark violations. Steinberger was since employed by Anthropic's rival OpenAI.)

Central to OpenClaw's enchantment was its functionality of permitting customers to have a persistent, private AI employee that they’ll message 24/7, at any time when they really feel like, over widespread messaging apps akin to iMessage, Slack, Telegram, WhatsApp and Discord, and have their AI message them again — not simply to talk with, however to carry out actual work for them by itself, from writing, sending and organizing e mail and information to creating complete functions, making use of for jobs on the consumer's behalf, to managing full ongoing social advertising and marketing campaigns. When the AI finishes a job, it may instantly alert the human consumer over their most well-liked messaging platform.

However OpenClaw additionally got here with a excessive diploma of safety threat (because it could possibly be given entry to a consumer's arduous drive and file system, or different private info, and run amok) and issue for non-technical customers, inspiring a wave of offshoots promising larger ease and safety, together with NanoClaw, KiloClaw and Nvidia's not too long ago introduced NemoClaw.

By giving Claude Code this similar primary performance — the flexibility for customers to message it from common third-party apps Discord and Telegram, and have it message them again when it finishes a job — Anthropic has successfully countered OpenClaw's enchantment and supplied one thing it doesn’t: the Anthropic model identify with its dedication to AI safety and security, and ease of use proper out of the field for much less technically inclined customers.

Know-how: The Bridge of the Mannequin Context Protocol

On the coronary heart of this replace is the Mannequin Context Protocol (MCP) open supply commonplace that Anthropic launched again in 2024. Consider MCP as a common USB-C port for AI: it supplies a standardized method for an AI mannequin to hook up with exterior knowledge and instruments. Within the new "Channels" structure, an MCP server acts as a two-way bridge.

When a developer begins a Claude Code session with the --channels flag, they aren't simply opening a chat; they’re spinning up a polling service.

Utilizing the Bun runtime—identified for its excessive pace in executing JavaScript—Claude Code displays particular plugins (at present Telegram and Discord).

When a message arrives, it’s injected immediately into the lively session as a <channel> occasion. Claude can then use its inside instruments to execute code, run checks, or repair bugs, and reply again to the exterior platform utilizing a specialised reply device.

The technical achievement right here is persistence. In contrast to a typical web-chat that instances out, a Claude Code session can now run in a background terminal or a persistent server (like a VPS), ready for a "ping" to spring into motion.

arrange Claude Code Connectors on Telegram and Discord

Establishing these native connectors requires Claude Code v2.1.80 or later and the Bun runtime put in in your desktop PC or Mac. Observe the directions right here or under.

1. Establishing Telegram

1. Create your Bot: Open BotFather in Telegram and use the /newbot command to generate a novel bot and entry token.

2. Set up the Plugin: Inside your Claude Code terminal, run: /plugin set up telegram@claude-plugins-official

3. Configure the Token: Run /telegram:configure <your-token> to save lots of your credentials.

4. Restart with Channels: Exit Claude and restart utilizing the channel flag: claude --channels plugin:telegram@claude-plugins-official

5. Pair your Account: DM your new bot on Telegram to obtain a pairing code, then enter it in your terminal: /telegram:entry pair <code>

2. Establishing Discord

1. Create an Utility: Go to the Discord Developer Portal, create a "New Utility," and reset the bot token to repeat it.

2. Allow Intents: Within the Bot settings, you should allow Message Content material Intent below "Privileged Gateway Intents."

3. Set up and Configure: In Claude Code, run /plugin set up discord@claude-plugins-official adopted by /discord:configure <your-token>.

4. Launch and Pair: Restart with claude --channels plugin:discord@claude-plugins-official. DM your bot on Discord and use the /discord:entry pair <code> command to complete the hyperlink.

Product: From Desktop to "In all places"

The instant sensible impression is the democratization of cellular AI coding. Beforehand, if a developer needed to test a construct standing or run a fast repair whereas away from their desk, they needed to depend on advanced self-hosted setups like OpenClaw.

With Channels, the setup is native. A developer can create a Telegram bot by way of BotFather, hyperlink it to Claude Code with a /telegram:configure command, and "pair" their account with a safety code. As soon as configured, the cellphone turns into a distant management for the event setting.

The product additionally introduces a "Fakechat" demo—a local-only chat UI that permits builders to check the "push" logic on their very own machine earlier than connecting to exterior servers. This displays Anthropic’s cautious, "analysis preview" strategy, making certain builders perceive the circulation of occasions earlier than exposing their terminal to the web.

Licensing: Proprietary Energy on Open Requirements

The licensing implications of this launch spotlight a rising pattern within the AI business: proprietary engines operating on open tracks. Claude Code stays a proprietary product tied to Anthropic’s business subscriptions (Professional, Max, and Enterprise).

Nevertheless, by constructing on the open-source Mannequin Context Protocol, Anthropic is encouraging a developer ecosystem to construct the "connectors" that make their mannequin extra helpful.

Whereas the core Claude "mind" is closed, the plugins for Telegram and Discord are being hosted on GitHub below official Anthropic repositories, possible permitting for group contributions or forks.

This technique permits Anthropic to keep up the safety and high quality of the mannequin whereas benefiting from the speedy innovation of the open-source group—a direct problem to the "free" however typically fragmented nature of purely open-source agent frameworks.

And since it's constructed on MCP, the group can now construct "Connectors" for Slack or WhatsApp themselves, relatively than ready for Anthropic to ship them.

Group Reactions: 'The OpenClaw Killer'

The response from customers, particularly AI observers on X, was swift and definitive. The sentiment was finest captured by Ejaaz (@cryptopunk7213), who famous that Anthropic’s pace of delivery—incorporating texting, hundreds of MCP expertise, and autonomous bug-fixing in simply 4 weeks—was "fucking loopy."

For a lot of, this replace renders local-first agent frameworks out of date. BentoBoi (@BentoBoiNFT) noticed, "Claude simply killed OpenClaw with this replace. You not want to purchase a Mac Mini. I say this as somebody who owns a one lol," referring to the widespread observe of builders shopping for devoted {hardware} to run open-source brokers like OpenClaw 24/7. By shifting this persistence into the Claude Code setting, Anthropic has simplified the "{hardware} tax" for autonomy.

AI YouTuber Matthew Berman summarized the shift succinctly: "They've BUILT OpenClaw."

The consensus amongst early adopters is that Anthropic has efficiently internalized probably the most fascinating options of the open-source motion—multi-channel help and long-term reminiscence—whereas sustaining the reliability of a tier-one AI supplier.

Whereas Anthropic’s Claude has lengthy been a favourite for its reasoning, it remained a "mind in a jar"—a stateless entity that waited for a consumer to kind earlier than it may suppose. In the meantime, open-source tasks like OpenClaw thrived by providing "always-on" persistence, permitting builders to message their AI from Telegram or Discord to set off advanced workflows.

Now, with Anthropic closing the hole, it's as much as the customers to decide on which strategy is finest for them.