Blog
Engineering insights on deploying, managing, and scaling AI agents in production.
One agent per user: how penpal is built on Maritime
penpal gives every learner a private OpenClaw agent on Maritime: 24 personas, one call to provision, and a friendship that lives on the agent's own disk.
Your agent's filesystem is now an API
The Maritime Files API: eight endpoints and both SDKs to browse, upload, download, and edit an agent's disk over HTTP, even while the agent sleeps.
Every agent framework is secretly the same five pieces
Strip the branding off OpenClaw, Hermes, ZeroClaw, or Claude Code and the same anatomy appears: a loop, a config, a state directory, a front door, a clock.
Running Claude Code on Maritime: The Step-by-Step Guide
Give Claude Code its own always-on computer in the cloud. Deploy the template, paste your Anthropic key, and message a real coding agent from anywhere.
Connect Your OpenClaw Agent to Hundreds of Apps with Composio
Composio handles OAuth, OpenClaw speaks MCP natively, and your Maritime agent gets Gmail, Slack, and Notion tools in minutes. Exact commands included.
CLI 1.5: a directory is an agent
The Maritime CLI grew a project model, one-command public web apps, scoped API keys, and a JSON contract designed for AI agents to drive.
ZeroClaw: how small can an agent get?
A full agent runtime in one static Rust binary. No interpreter, no dependency tree, and a footprint that makes snapshots small and wakes fast.
Hermes: the agent with a front door
A messaging-native agent you can talk to from the CLI or the dashboard, whose channels have a pairing wall: strangers get held at the door until you approve them.
Traditional Hosting Is Broken for AI Agents
VMs, containers, serverless: none of it was designed for AI agents. Here is why every hosting model fails them, and what agent-native infrastructure looks like.
Security Best Practices for Production AI Agents
AI agents execute code, call APIs, and handle sensitive data. These are the security practices that actually matter in production.
Why AI Agents Need Sleep/Wake Architecture
Most AI agents sit idle 99% of the time, burning money on always-on infrastructure. Sleep/wake architecture changes that. Here is how it works and why it matters.
Building Event-Driven AI Agents with Webhook Triggers
Stop polling. Webhook triggers let your AI agents react to real-world events instantly, from Stripe payments to GitHub pushes to Slack messages.