OpenClaw has released its version 2.0, which the project describes as its biggest update since the launch. The open-source AI agent platform says the release was built with contributions from 933 developers, including 569 people making their first contribution and includes more than 16,000 pull requests. The update brings changes across installation, messaging, memory, skills, models, automation, browser functionality, native apps, plugins and security.
OpenClaw 2.0 offers a revamped installation process aimed at getting new users up and running faster. The platform can now make use of resources users may already have on their computers, including ChatGPT or Claude subscriptions, API keys and locally installed models. OpenClaw has also reduced the amount of configuration required during the initial setup, with users able to complete more of the process by interacting with their AI agent.
The browser interface has also been rebuilt as a central part of the OpenClaw experience. Users can start conversations, continue existing tasks and monitor their Claw’s activity from the redesigned interface. The project says the changes are intended to make it easier for users to start with a simple automated workflow and gradually expand what their agent can handle.
OpenClaw 2.0 also introduces shared cloud sessions, allowing multiple people to work with the same AI agent while retaining its existing context. The feature was developed after the OpenClaw team began using agents more extensively during the development of the release and needed a way to share ongoing tasks without starting over.
The project says the new multiplayer functionality can allow users to bring colleagues or family members into an active workflow or hand tasks between people while preserving the agent’s context.
OpenClaw said the latest release includes roughly half of all pull requests ever merged into the project. The team delayed the release to ensure the update worked for both new installations and existing users upgrading from earlier versions.