Notifly use cases — ready-made recipes

Notifly is a universal push-notification channel for everything that matters in your work. Instead of opening dashboards, scrolling through email and checking statuses by hand, you get an instant message on your phone or in the browser — and anything left unread Notifly re-sends by email, SMS, a phone call or to your team chat.

This section collects practical, copy-paste-ready usage examples. Some recipes are set up entirely in the admin panel with no code; the rest boil down to a single HTTP call:

curl -X POST "https://api.notifly.ru/message?token=APP_TOKEN" \
     -H "Content-Type: application/json" \
     -d '{"title":"Title","message":"Text","priority":5}'

What do you want to do?

• Find out that a website or server stopped responding → Active monitors. You need: a public address of the site or port. No code — the monitor is created in the admin panel.
• Find out that a cron job or backup did not run → Heartbeat for cron and backups. You need: access to the crontab or systemd timer on the server — to add one curl line.
• Get errors from your website → Frontend errors or backend errors. You need: for the frontend — paste a ready-made <script> from the admin panel into the site’s <head>; for the backend — add middleware to your Node.js, Python or Go code.
• Get leads and orders from a CMS website → Tilda, WordPress and other CMS. You need: for Tilda — just the form settings and a Webhook Router rule, no code; for WordPress, 1C-Bitrix and the other CMS — put a PHP file on the site’s server, so you need access to its files and at least a basic understanding of PHP.
• Get Prometheus, Grafana or Zabbix alerts on your phone → Prometheus and Grafana, Zabbix. You need: admin rights in those systems — to add a receiver to alertmanager.yml, a contact point in Grafana or a media type in the Zabbix UI. No programming required.
• Ask a person from a script or AI agent and wait for the answer → Human-in-the-loop and Questions in channels (Ask). You need: to add an HTTP request to /ask to your script or agent — ready-made examples with curl and Python.
• Personal reminders on your phone → Reminders and Pomodoro. You need: to copy commands into a terminal — at and cron on Linux and macOS or PowerShell on Windows.

Recommended for a start

Four recipes that are the easiest to begin with: each one pays off on day one.

1. Active monitors — a push when a site, API or port stops responding, and another one when it is back. No code, everything is set up in the admin panel. Requirement: the target is reachable from the internet.
2. Heartbeat for cron and backups — Notifly tells you when a scheduled job did not report in time. Add && curl … to the crontab line.
3. Frontend errors — your visitors’ JS errors arrive as push notifications with the stack trace and URL. Pasting a <script> into the site template is enough, no backend needed.
4. Human-in-the-loop — an agent or script asks a question with “approve / deny” buttons and waits, while you answer right from the notification. Add a request to /ask to your code (a Python example is ready).

So that nothing important gets lost, connect an escalation chain to any of them: Notifly re-sends an unread push by SMS, a phone call or to your team chat. Set up in the admin panel, no code.

By role

🛠 System Administrator

Monitoring servers, disks, services, security — everything that should “ring” when something goes wrong. Ready-made scripts for Linux, systemd, cron, Fail2ban, SSL, plus receiving alerts from Prometheus Alertmanager, Grafana and Zabbix.

💻 Developer / Vibecoder

How to embed notifications in frontend and backend code: production errors, completion of long-running tasks, CI/CD events, successful deploys. Examples in Node.js, Python, Go and React.

🤖 AI agents and LLM apps

Asking a person a question with buttons straight from an agent, completion and stalls of long tasks, LLM API spend, provider availability, answer quality, RAG and security.

🌐 From CMS sites

WordPress, 1C-Bitrix, Drupal, MODX, Joomla, OpenCart, PrestaShop and Tilda — recipes for popular CMS and site builders: new comments, orders, form leads, registrations, PHP errors.

🧑‍💼 Personal User

Notifications from your home PC or laptop: low battery, VPN disconnect, completion of a long download, reminders, network events.

---

Basic setup

All examples assume that you:

1. Have registered at Notifly (app.notifly.ru).
2. Created a channel in the admin panel and obtained an app token (prefix A).
3. Opened the feed in the web interface or installed the Android app (APK from the website). You can also send from the terminal with the CLI.

Save the token in an environment variable so you don’t have to include it in every example:

export NOTIFLY_URL="https://api.notifly.ru"
export NOTIFLY_TOKEN="AGdjfk_L.dKe8q"

And then the minimal send becomes:

curl -s "$NOTIFLY_URL/message?token=$NOTIFLY_TOKEN" \
     -H "Content-Type: application/json" \
     -d '{"title":"Test","message":"It works!"}'

The token can also be passed in the X-Notifly-Key: $NOTIFLY_TOKEN header — that way it doesn’t end up in proxy logs or your shell history.

Priorities

In Notifly each message has a priority (priority, a number 0–10). Clients interpret it differently:

Priority	Behavior in Android client
0–3	Silent, no sound
4–7	Normal notification
8–10	Loud, heads-up

Use high priority only for truly critical events.

So that alerts don’t wake you for nothing

Three settings that come in handy in almost every recipe:

• Delivery & escalation — if a push isn’t read within N minutes, the message goes out by email, SMS, a phone call, to Telegram, Slack, Discord, Microsoft Teams, Mattermost, PagerDuty or Opsgenie. Once it’s read, escalation stops.
• Quiet hours & digest — at night only urgent messages (priority at or above the threshold) are escalated; the rest arrives in the morning as a single summary.
• Maintenance windows — during planned work monitors and heartbeats stay silent and the status page keeps its uptime. See the recipe “Planned work without false alerts”.