OpenClaw ships 2.0, the largest update in its history
OpenClaw, the open source personal AI agent project, has released OpenClaw 2.0, which its team calls by far the largest update in the product's history. The release is composed of over 16,000 pull requests contributed by 933 people, 569 of them first-time contributors, and touches installation, messaging, memory, skills, models, automations, the browser and native apps, plugins, security and what the team describes as a very long tail of fixes.
Before this release, OpenClaw had shipped 106 releases in 230 days, most of them within a day or two of the previous one. Going nearly seven weeks without a release was unusual for the project. The team says the slower shipping cadence was not a sign of slower development: the team itself was growing, and the rising volume and pace of work outgrew both OpenClaw's underlying foundation and the process used to ship it, so both were reworked at the same time. The team also says this single release contains roughly 50% of every pull request ever merged into OpenClaw, and that the extra time went into making sure it worked both for people installing OpenClaw from scratch and for people upgrading an existing Claw, rather than shipping fast and breaking what existing users already had.
On installation, first-time setup now starts from what is already on a person's computer, including existing ChatGPT or Claude subscriptions, API keys and local models. The team says it cut or simplified a lot of configuration and moved the remainder out of initial setup, so people reach a first conversation faster and can finish configuring their Claw by talking to it directly. The browser app, described as where most people meet OpenClaw for the first time, was rebuilt as what the team calls a first-class experience for setup, ongoing work and watching a Claw work live.
The release also adds shared cloud sessions, letting more than one person join a running Claw session, collaborate on a task or hand it off entirely without losing what the Claw already knows about the task. The team says OpenClaw previously had no way to bring a second person into ongoing work without losing that context, and that this multiplayer capability is now how the team itself works while building OpenClaw.
The team frames the update around use cases that scale from simple to more involved: an example given is a Claw that watches an inbox for a child's school emails and forwards anything important to Telegram, then, from that same setup, is asked to check an iMessage question against a saved email receipt and reply on the sender's behalf. The post closes by stating OpenClaw is open source and belongs to the people who use and build it, and that the team is not asking anyone to trust one company, model or AI provider with it.
Key facts
- OpenClaw 2.0 is built from over 16,000 pull requests contributed by 933 people, including 569 first-time contributors, which the OpenClaw team calls the largest update in the product's history.
- It follows 106 releases shipped in 230 days at a near-daily cadence; this release took nearly seven weeks longer because the team grew and reworked both OpenClaw's foundation and its release process at the same time.
- The team says this one release contains roughly 50% of every pull request ever merged into OpenClaw, and that the extra time was spent making sure it did not break existing installs while upgrading.
- Installation is simplified: first-time setup now starts from what is already on a user's machine, including existing ChatGPT or Claude subscriptions, API keys and local models, and the browser app was rebuilt as the primary entry point.
- A new shared cloud sessions feature lets multiple people join or hand off the same running Claw session without losing its accumulated context; the team says it already uses this internally to build OpenClaw.
Why it matters
OpenClaw 2.0 is not a routine point release. The team's own numbers, over 16,000 pull requests from 933 contributors, roughly half of everything ever merged into the project, mark a release large enough that the team chose to slow its usual near-daily shipping cadence rather than ship it in pieces. That choice signals the update reworks OpenClaw's underlying foundation and release process, not just its feature list, which is why installation, the browser app, and how a Claw is shared between people all changed together rather than separately.
Who it affects
People running OpenClaw already are the most directly affected: the team says it deliberately spent extra time so upgrading an existing Claw would not break what was already configured. New users benefit from the simplified, more automatic installer. Anyone who wants to share a running agent with a family member or teammate, rather than start a separate one from scratch, gains the new shared cloud sessions capability. The 933 contributors, more than half of them contributing for the first time, are also directly credited as the people who built the release.
How to use it
OpenClaw 2.0 is available now as an open source release. First-time installation starts from whatever is already on a person's computer, existing ChatGPT or Claude subscriptions, API keys, or local models, with most configuration moved out of initial setup and completable by talking to the Claw afterward. The browser app is the main place to set things up, watch a Claw work, or return to a task in progress. Shared cloud sessions let a second person be brought into a live task or have it handed off to them with context intact, which the team presents as useful for families or teams splitting work.
How solid is it
The scale claims come from the team's own release post rather than an independent audit: 933 contributors, 569 of them first-time, over 16,000 pull requests, and a prior cadence of 106 releases in 230 days. These are specific, checkable figures rather than vague marketing language, and the project's history of near-daily shipping before this release suggests an active, established codebase rather than a one-off announcement. No breakdown of the 16,000-plus pull requests by feature versus fix versus refactor is given, and no independent source confirms the contributor or pull request counts.
Risks and caveats
Everything here comes from OpenClaw's own launch post; there is no third-party verification of the contributor count, the pull request total, or the claim that upgrades will not break existing installs. No calendar date is given for the release, only relative terms like 'nearly seven weeks' and 'nearly two months.' A release this large, by the team's own account roughly half of everything ever merged into the project, carries an inherent risk of regressions for existing users even with the stated extra testing time, and that risk cannot be assessed from the announcement alone.
“We started by simplifying installation and rebuilding the browser app as a first-class experience, but doing that properly meant carrying the cleanup through the rest of OpenClaw until it became OpenClaw 2.0.”
— OpenClaw team, release announcement