OpenClaw 2.0 goes multiplayer

OpenClaw is an open-source AI agent platform designed to automate your everyday tasks and connect your workflows, handling everything from email monitoring to cross-app actions. On August 30, it dropped version 2.0, by far the biggest release in its history, packed with over 16,000 pull requests from 933 contributors, including 569 first-timers.

Before this update, the project had shipped 106 releases in 230 days, usually pushing code every day or two. But the rapid growth forced the team to pause for nearly seven weeks to overhaul the platform from the ground up, rewriting roughly 50% of all PRs ever merged into the codebase.

What it can do

OpenClaw starts with simple automations and scales as you need it to. A basic workflow might just monitor your inbox for your kids' school emails and send a summary to Telegram whenever important homework is due.

From there, tasks can reach across multiple apps without added friction. If a family member texts asking which tablet you bought for your dad, you can simply tell your Claw to look up the receipt in your email and send the reply. The agent handles the retrieval and messaging automatically.

The browser app has been completely rebuilt as a first-class experience, letting you manage your setups, return to running workflows, or watch the agent execute tasks live.

Setup and multiplayer workflow

Setting up self-hosted agents used to be a headache, but OpenClaw 2.0 slashes initial friction. It detects what is already on your machine, letting you hook in existing ChatGPT or Claude subscriptions, API keys, or local models right out of the box.

Heavy configuration menus are gone. You get into a conversation immediately and finish customizing your agent simply by chatting with it.

The real game-changer is multiplayer shared cloud sessions. Previously, bringing a collaborator into an agent's workflow meant losing the existing context. Now, teams can jump into the same session with memory intact.

Founder Peter Steinberger noted that the team built 2.0 using this exact setup, remarking that traditional local harnesses feel like relics of the past. The project is fully open source, giving you full ownership without lock-in to any single vendor.

Why it matters

If you have been holding off on personal AI agents because setup felt too fragile or technical, OpenClaw 2.0 is worth checking out now. Being able to ride on existing consumer subscriptions or run local models keeps costs predictable and privacy in your hands.

Adding multiplayer capabilities pushes autonomous agents past solo tinkering into real team collaboration. When agents can retain full context while being passed between humans, they start functioning like genuine digital teammates.