Notifly for the developer (vibecoder)

If you write code alone or in a small team, you almost never have a full monitoring system. Notifly fixes that in 5 lines of code: any important thing in your application becomes a push notification.

When this is especially useful

• Pet project in production. You don’t want to spin up Sentry/Datadog for 100 RPS, but you still want to know about 500s.
• Long-running tasks. Model training, dataset assembly, video rendering — repeatedly poking ssh gets tedious.
• Deploy. You want to feel the flow: “deploy started” → “reached production” → “smoke tests OK”.
• Background jobs. Cron job, queue worker, scheduled task — anything that runs “out there”.

Where to start

1. Frontend errors — paste the ready-made WebScript “Console errors” snippet from the admin panel into your site’s <head>. No backend or proxy needed.
2. Backend errors — middleware for Express, FastAPI or net/http: add it to your code and put the token into environment variables.
3. CI/CD events — a step with curl in GitHub Actions or GitLab CI. Requires access to the repository secrets.
4. Long-running task completion — copy the bash wrapper and run any long command through it: you get a push on success or failure.

Coding with AI agents or building an LLM app? See the separate section AI agents and LLMs: questions to a person with buttons, stuck agents, token spend.

Available recipes

Scenario	Stack
Backend errors and events	Node.js, Python, Go
Frontend errors	React, vanilla JS
CI/CD events	GitHub Actions, GitLab CI
Deploy notifications	Docker, k8s, any PaaS
Completion of long tasks	Python, ML, data processing
New signups and first sales	Node.js, Python, cron
Payments and payment failures	Node.js, Python, Stripe, YooKassa
Background job failures	Celery, BullMQ, Sidekiq
Anomalous traffic and API abuse	Node.js, Python
Contact form leads	Node.js, Python
AI agents and LLMs	LLM API, AI agents, RAG, MCP, eval

Basic wrapper

In any language it’s realistically two lines. For example, in Python:

import os, requests

def notify(title, message, priority=5):
    requests.post(
        f"{os.environ['NOTIFLY_URL']}/message",
        params={"token": os.environ["NOTIFLY_TOKEN"]},
        json={"title": title, "message": message, "priority": priority},
        timeout=5,
    )

Below on each page — ready recipes for specific tasks.