Monitoring API for AI agents · MCP + OpenAPI 3.1 · Enterno.io

Anatoly Oshmanovsky

Category: Monitoring for AI workflows

The monitoring API built for AI agents.

Production-grade uptime, SSL, DNS and security checks for every domain — exposed as a Model Context Protocol server, an OpenAPI 3.1 spec and an llms.txt catalog. Your AI agent can discover it, authenticate, and start running checks in under a minute.

Get API key — free → Read API docs

MCP server-card A2A agent-card OpenAPI 3.1 llms.txt

Why AI agents specifically?

🔌

Native MCP server

Point Claude Desktop, Cursor, Zed, or any MCP client at https://enterno.io/mcp. No wrapper, no shim — the checks are first-class MCP tools with JSON schemas and auth.

📇

Full discovery stack

We ship every well-known descriptor an agent looks for: ai-plugin.json, Google A2A agent-card.json, RFC 9727 api-catalog, OpenID config, HTTP-signatures directory. No guessing.

🧠

LLM-friendly context

A daily-regenerated llms.txt catalog gives ChatGPT, Perplexity, Google AI Overviews and Bing Copilot machine-readable page descriptions — so your agent cites us correctly on the first try.

🔑

Agent-safe auth

Per-key scopes (check, dns, ssl, ping, ip, monitors, webhook), idempotency keys on POST, IP allowlists, rotatable tokens. Give your agent the minimum it needs.

Three ready-to-build agents

Case 1 · DevOps

Autonomous monitoring agent

Claude/Cursor, given the MCP server, auto-creates monitors for every service it finds in your repo docker-compose.yml, wires them to Slack, and sets SSL-expiry alerts 14 + 3 days out. No dashboard clicks.

@enterno list-monitors
@enterno create-monitor url=https://api.example.com type=http
@enterno webhook-add events=monitor.down,ssl.expiring url=https://hooks.slack.com/...

Case 2 · SRE

AI-driven incident responder

PagerDuty fires. Your agent queries /api/v4/events for the 30-day history, cross-references DNS + SSL + WHOIS state, drafts a postmortem, and opens a Jira ticket with the root-cause hypothesis. Engineer just reviews.

@enterno events --webhook_id=42 --event_type=monitor.down --since=24h
@enterno check_dns domain=example.com
@enterno check_ssl domain=example.com
# → agent writes postmortem + opens Jira ticket

Case 3 · SEO / content

SEO automation agent

n8n / LangGraph pipeline runs a weekly crawl of your top pages through the security + HTTP-header + schema tools. When a competitor drops their HSTS or breaks their schema — your agent drafts a blog post about it.

@enterno check_http url=https://competitor.com
@enterno security-scan url=https://competitor.com
# → n8n branches: if hsts.missing → draft_post.md

Capability catalog

Every tool below is also published as machine-readable JSON Schema at /.well-known/ai-capabilities.json . Filter, expand, copy a sample call — or run one live below.

Try it in your browser

The try-it console is for logged-in users. Your API key never leaves the server — we proxy the request.

Sign up — free → Log in

Pick your agent

Same key, three transport flavours. Pick the one your AI runtime supports — they all hit the same MCP endpoint.

Claude Desktop · Anthropic — native MCP, streamable HTTP

Get a key. Sign up, visit Settings → API Keys, copy your ent_… key.

Edit claude_desktop_config.json:

{
  "mcpServers": {
    "enterno": {
      "url": "https://enterno.io/mcp",
      "transport": "streamable-http",
      "auth": { "type": "apiKey", "header": "Authorization",
                "prefix": "Bearer", "value": "ent_your_key" }
    }
  }
}

Restart Claude Desktop, then ask: "Check SSL on stripe.com. Alert me if it expires in <14 days."

Cursor IDE — MCP plugin, same config

Cursor → Settings → MCP → Add server.
Paste the same JSON (as Claude Desktop above). Cursor uses identical MCP transport — no separate adapter needed.
@enterno in any chat to invoke checks while you code.

Direct API (no MCP) — curl / Python / any language

Single SSL check via REST:

curl -H "X-API-Key: ent_your_key" \
  "https://enterno.io/api/v4/ssl?host=stripe.com"

Full spec: /swagger · /api/docs.
Prefer LangChain / LangGraph? Pull our agent-card from /.well-known/agent-card.json — Google A2A-compatible.

5 prompts to try right now

Each one resolves to a single tool call. Copy any into Claude Desktop / Cursor with the MCP server connected.

SSL expiry triage: "List the SSL expiry dates for stripe.com, github.com, and aws.amazon.com. Sort by remaining days, ascending."
DNS drift detection: "Compare TXT records for enterno.io across 1.1.1.1, 8.8.8.8, and 77.88.8.8. Highlight any divergence."
Security regression after deploy: "Run a security-headers scan on staging.example.com. List anything that dropped vs grade A."
Multi-region uptime check: "Ping example.com from EU and US. Tell me if response-time differs by more than 200ms."
AI visibility audit: "Check whether ChatGPT, Claude, and Perplexity cite stripe.com when asked about payment APIs. Summarise the gap."

Ready to build?

Free tier: 5 monitors, MCP access, 50 API calls/day. Upgrade when your agent needs more.

Create free account →