Enhance browser automation with modal dialog handling

What changed

OpenClaw added modal dialog support to its browser automation features on 21 May 2026. The tool now includes pending and recently handled dialogs directly in page snapshots. It also returns a blockedByDialog status flag and accepts dialog answers through the existing CLI interface.

These updates apply to the self-hosted agent running on a local VPS. No price changes were announced. Users still pay only for their VPS and LLM tokens.

Why it matters

Modal dialogs frequently block automated browser flows on login pages, cookie banners, and confirmation prompts. The new status and snapshot data let the agent detect blocks without custom workarounds. This reduces failed runs on common sites that vibe builders automate daily.

The change strengthens OpenClaw's position against closed browser agents that already handle dialogs. It keeps the tool viable for solo operators who rely on persistent, self-hosted automation rather than switching to managed services.

How to use it

Update to the latest OpenClaw release via the CLI command documented on the project GitHub. Enable browser control in your YAML config and run any task that triggers dialogs. Check the snapshot output for the new blockedByDialog field and respond with the answer-dialog command when needed.

The feature works on any connected LLM provider. Test first on a non-critical site to confirm dialog detection before adding it to production workflows.

Watch for

Confirm the bet if dialog handling reduces failed browser tasks by at least 30 percent in user reports over the next month. The bet breaks if snapshot size grows too large and slows down heartbeat cycles. Expect the next move to be similar handling for file upload dialogs and JavaScript alerts.