| name | ohdear |
| description | Manage Oh Dear website monitoring using the ohdear CLI. ALWAYS activate when the user's message contains an ohdear.app URL (e.g. https://ohdear.app/monitors/123/check/broken-links/report). Extract the monitor ID and check type from the URL and use the CLI to fetch the data. Also use when the user wants to check if their site is up, view uptime or downtime, list or manage monitors, view broken links, mixed content, or certificate health, manage maintenance periods or status pages, configure cron checks, view DNS history or DNS blocklists, check port scan results, view sitemap checks, check application health, view Lighthouse performance reports, view crawled URLs summary and details, manage tags, manage notification destinations, manage teams, or interact with ohdear.app in any way. |
| license | MIT |
| metadata | {"author":"ohdearapp","version":"5.2.0"} |
Oh Dear CLI
The ohdear CLI manages Oh Dear website monitoring from the terminal. Every Oh Dear API endpoint has a corresponding command.
Prerequisites
ohdear --version
composer global require ohdearapp/ohdear-cli
Authentication
ohdear login
ohdear logout
If any command returns 401, run ohdear login again.
Command naming
Commands use kebab-case names exactly as listed in references/commands.md. Do NOT guess command names — always look them up. Common mistakes to avoid:
sites:list → list-monitors
monitors:get → get-monitor
uptime:get → get-uptime
Always check references/commands.md before running a command. If unsure, run ohdear <command> --help to verify parameters.
Output
Commands output human-readable text by default. For any task that requires analysis, comparison, filtering, or counting, always use --json to get structured output you can parse:
ohdear list-monitors --json
ohdear get-uptime --monitor-id=1 --json
Other format options: --yaml, --minify. Use -H to include response headers.
Note: --json works on API commands (list-, get-, etc.), not on utility commands (login, logout, clear-cache).
JSON structure: --json responses are paginated objects {data: [...], links: ...}, not bare arrays. Always use .data[] to access items (e.g. jq '.data[]'), never .[].
Teams
API keys are user-scoped. A user may belong to multiple teams. Monitors have a team_id field but no team_name. To resolve team names to IDs:
ohdear get-me --json | jq '.teams[] | {id, name}'
Then filter monitors by team_id. Do NOT use list-managed-teams — that's a reseller-only endpoint.
When presenting results to the user, summarize clearly:
- Monitors: Table with ID, URL, friendly name, status, checks enabled
- Uptime: Percentage uptime and downtime periods with start/end times
- Broken links: URLs with HTTP status codes and source page (use
--run-id with historical run IDs to compare across runs — see workflows)
- Crawled URLs: Always start with
get-crawled-urls-summary (tiny payload: totals + by-type breakdown). Only use list-crawled-urls-details with jq filters — it returns 100 items/page and can be very large. Never dump raw details into context.
- Certificate health: Issuer, expiration date, and issues found
- Cron checks: Name, frequency, last ping time, status
- Status pages: Title, URL, associated monitors
- Lighthouse reports: Performance, accessibility, best practices, SEO scores
Documentation
Oh Dear docs are available in clean markdown for easy reading. Append .md to any docs URL or request with Accept: text/markdown:
https://ohdear.app/docs/general/introduction.md
https://ohdear.app/docs/api/monitors.md
Use this to look up feature details, API parameters, or setup instructions when helping the user.
Reference
- Full command list: See references/commands.md
- Step-by-step workflows: See references/workflows.md
- Monitor setup, downtime investigation, maintenance windows
- Broken link audits, certificate monitoring, status pages
- Cron checks, Lighthouse reports, application health, tags