API & MCP

Everything you can see in the dashboard, from scripts, CI pipelines and AI assistants. Available on every plan.

Machine-readable: the OpenAPI spec describes every endpoint, for client generators, Postman and coding agents.

Authentication

Create a token under Profile → API & data. Choose read-only or read and write, and optionally limit it to specific monitors. Send it as a bearer token:

curl https://observe.live/api/v1/monitors \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Accept: application/json"
  • Read-only tokens get 403 on POST, PATCH and DELETE.
  • Tokens limited to specific monitors get 404 for other monitors, and 403 for creating monitors and for organization endpoints.
  • Team permissions apply: you only see monitors your teams can see.
  • Requests are limited to 120 per minute per token. Responses are JSON; IDs are strings; times are ISO 8601 in UTC.
  • API tokens themselves are created and revoked only in the dashboard, on purpose: a leaked token cannot mint new tokens or keep itself alive.

Reference

  • API reference — every endpoint with its parameters, fields and examples: monitors and their check settings, results, incidents, maintenance, status pages, notification channels, members and teams, reports, organization settings and server monitors.
  • Reports: save or run a report across monitors (POST /organizations/{id}/reports/run): per monitor and per group uptime, downtime, incidents, MTTR, avg / p50 / p95 response time, expiry, Lighthouse, crawl health and DNS record changes (nameserver and DMARC, DKIM, SPF changes counted), filtered by tag or team. Monitor groups are managed under /organizations/{id}/groups; generate-report does the same over MCP.
  • MCP server — connect Claude or Cursor, and every tool with an example.
  • Check types — what each check does, how often it runs, and its settings.

Base URL: https://observe.live/api/v1

Quickstarts

Create a monitor

Find your organization ID, then post a URL. Uptime and SSL checks start straight away; the response is the monitor, with its ID.

curl https://observe.live/api/v1/organizations -H "Authorization: Bearer YOUR_TOKEN"

curl -X POST https://observe.live/api/v1/monitors \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"organization_id": "ORG_ID", "url": "https://example.com", "label": "Marketing site", "tags": ["marketing"]}'

Leave organization_id out to use your first organization. Over your plan's monitor limit the request is a 422 on url.

Enable all checks

Each check type has an *_enabled field, and any of its settings can go in the same request. Each one you turn on runs its first check now.

curl -X PATCH https://observe.live/api/v1/monitors/MONITOR_ID \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "uptime_interval_minutes": 1,
    "uptime_check_string": "Add to cart",
    "domain_monitoring_enabled": true,
    "dns_record_monitoring_enabled": true,
    "dns_blocklist_monitoring_enabled": true,
    "sitemap_monitoring_enabled": true,
    "robots_txt_monitoring_enabled": true,
    "lighthouse_monitoring_enabled": true,
    "lighthouse_thresholds": {"min_score": 80},
    "crawl_monitoring_enabled": true,
    "crawl_interval_minutes": 10080,
    "crawl_max_pages": 1000
  }'

Then read results per type, e.g. GET /monitors/MONITOR_ID/lighthouse. Every field is listed under check types.

Get alerts via webhook

Add a webhook destination (admins), choose its events, and send a test:

curl -X POST https://observe.live/api/v1/organizations/ORG_ID/notification-channels \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"type": "webhook", "label": "Incident bot", "webhook_url": "https://hooks.example.com/observe", "event_types": ["uptime_incident_opened", "uptime_incident_resolved"]}'

curl -X POST https://observe.live/api/v1/organizations/ORG_ID/notification-channels/CHANNEL_ID/test \
  -H "Authorization: Bearer YOUR_TOKEN"

On-call tools work the same way: type pagerduty with routing_key, or opsgenie / jsm (Jira Service Management, Opsgenie's successor) with api_key, priority and responder_team. See the Opsgenie and JSM guide.

Each alert is a JSON POST with an event_type and the organization, monitor and incident details. The Observe-Signature header is the HMAC-SHA256 of the raw body with the channel's signing_secret (in the create response, and on the Notifications page); compare it before you trust the request. A test has "test": true. Leave event_types out for every event; the event names are listed under createNotificationChannel in the API reference.

Use with Claude via MCP

Create a token (read-only to ask questions, read and write to act), then add the server:

claude mcp add --transport http observe https://observe.live/mcp \
  --header "Authorization: Bearer YOUR_TOKEN"

Then ask, for example: “Which monitors are down?”, “List the broken links from the last crawl”, “Schedule maintenance for the API tonight 22:00 to 23:00 UTC” or “Post a status update that we are investigating slow checkout”. It uses the same permissions as the dashboard.

Cursor, Claude Desktop and every tool

Heartbeats

Cron jobs report in without a token, using the monitor's heartbeat URL (heartbeat.ping_url in the API): GET or POST to it on success, or append /starting or /failed.