Usectl for agents

Connect your agent to usectl

Run open-source AI agents in network-isolated sandboxes with persistent storage, while built-in copilots triage failures and manage incidents safely.

https://manager.usectl.com/api/mcp

Choose your agent

Hermes Agent

Persistent memory & isolated networks

OpenClaw

Token-protected workspace & chat gateways

Task Agent

Background tasks with tools & web access

Built-in infrastructure copilots

Devo Admin Copilot

Propose-and-Approve workflows with zero direct writes. Approve from the CLI, the web panel, or the iOS app.

Failed-Deploy Triage

Auto-analyzes build logs on failure at zero user cost, commenting fixes on PRs.

Incident Remediation

Monitors Grafana alerts across notify, approve, or auto modes with 1-tap approvals.

Weekly Machine Digests

A weekly summary of each machine every Monday at 07:00 UTC: deploys, incidents and pod health.

Built for untrusted execution

Network-isolated agents

Internet-only egress, no access to your cluster or private networks.

Scoped API Keys (usectl tokens)

Least-privilege keys bound to machines, read or write access, and an expiry.

Explore templates

Hermes on usectl

OpenClaw on usectl

Task Agent

Secure AI & Autonomous OPS

Put your agents to work on usectl

Host open-source AI agents in network-isolated sandboxes and let built-in copilots manage your infrastructure.