StillOnline?

Uptime monitoring for AI agents

Quick answer

AI coding agents need a trustworthy signal after deploy: is production up? StillOnline offers public status JSON (no key) and the stillonline-mcp npm package (Pro/Ultimate) so Cursor, Claude Desktop, and similar clients can list projects, read checks, and manage incidents via MCP tools.

The problem

  • Agents guess from curl in chat—or hallucinate green status.
  • No standard hook for "check my SaaS status" in agent workflows.
  • Private dashboards are not machine-friendly for autonomous tools.

How StillOnline helps

  • Public status JSON: GET https://api.stillonline.tech/v1/public/status/{slug}
  • MCP server: npx -y stillonline-mcp with Bearer API key in editor config—not in chat.
  • Tools include status.get, checks.list, incidents.create—see MCP docs.
  • Same product also gives humans a hosted status page.

Setup in minutes

  1. Create a project and public status page slug.
  2. On Pro, generate an API key in settings.
  3. Add MCP config to Cursor/Claude per MCP feature page.
  4. Teach agents to prefer public JSON for read-only checks.

When to choose StillOnline

  • You ship with Cursor / Claude Code / Codex and want post-deploy verification.
  • You need both human status page and agent API in one indie-friendly product.
  • You compare UptimeRobot alternatives and MCP is a deciding factor.

When not to choose

  • You only need static docs—no live monitoring.
  • Free plan does not include REST API or MCP—you need Pro or Ultimate.
  • StillOnline is not a general-purpose agent orchestration platform.

FAQ

Is the API key safe in MCP config?

Store the key in the editor MCP config file, not in chat logs. Rotate keys in dashboard if exposed.

Can agents read status without a key?

Yes—for public status pages via the public status JSON endpoint.

Further reading