Amazon Web Services announced the general availability of OpenClaw on Amazon Lightsail on March 4, 2026, making it possible to deploy a personal AI agent on AWS infrastructure with a single click.
How It Works
Users select OpenClaw as a Lightsail blueprint, choose an instance size (AWS recommends at least 4 GB memory), and launch. The instance comes pre-configured with Amazon Bedrock as the default AI model provider, using Anthropic Claude Sonnet 4.6 out of the box. After launch, users pair their browser session using a gateway token displayed in the SSH welcome message.
AWS CloudShell automates IAM role creation and Bedrock API access policy attachment, removing the manual setup that previously made self-hosted OpenClaw on AWS cumbersome.
Messaging Channel Support
Once deployed, the Lightsail OpenClaw instance can connect to WhatsApp, Telegram, Slack, Discord, Microsoft Teams, and other messaging platforms — the same channels supported by any OpenClaw installation. Users can also interact through the built-in web gateway.
Security Recommendations
AWS explicitly recommends against exposing the OpenClaw gateway to the open internet, advising users to rotate authentication tokens frequently and store secrets in environment files rather than hardcoding them. The announcement noted that the launch follows documented vulnerabilities including CVE-2026-25253, underscoring the importance of keeping installations updated.
Available Everywhere
OpenClaw on Lightsail is available in all AWS commercial regions where Lightsail operates, giving users worldwide access to a managed OpenClaw deployment option backed by AWS infrastructure.
For users who prefer a simpler, fully managed experience without AWS configuration, OpenClaw Launch offers one-click deployment with automatic updates and built-in monitoring.