n8n Telegram Automation: Alerts, Bots, and Notification Workflows

Telegram Gets Read. Email Doesn't.

Most alert pipelines route to email. That's fine until the alert actually matters — a service down at 2 a.m., a high-value user hitting a plan ceiling, a negative review going live. Email sits unread for hours. Telegram delivers in under two seconds. No spam filter to fight. The phone buzzes.

n8n Telegram automation makes Telegram the last step in any workflow: monitoring jobs, community scanners, sales alerts, ops pipelines. The n8n-nodes-base.telegram v1.2 node handles both sending and receiving, and you can wire your first bot in about four minutes.

This post covers the full setup, three concrete implementation patterns, and named templates you can deploy in under ten minutes.

What You Can Automate with n8n and Telegram

Telegram fits wherever an alert needs to interrupt rather than just inform:

• Uptime monitoring — ping API endpoints every 15 minutes, fire a Telegram message the moment response time crosses your latency threshold or a 5xx comes back
• Lead scoring alerts — route AI-scored hot leads to a sales group the moment they qualify, complete with post link, intent score, and a ready-to-post reply draft
• Community signal scanning — monitor subreddits for high-intent buyer posts, score them 1–10 with AI, push ranked signals to Telegram before a competitor sees them
• SaaS usage thresholds — notify the sales team when a free user hits 95% of their plan limit, flagging the right window for a personal conversion call
• Negative review interception — fire a Telegram alert the moment a low-star review arrives, so support can respond before it accumulates views overnight
• Scheduled digest summaries — push a daily Telegram summary of open tickets, queue depth, or pipeline value so teams stay informed without a dashboard login

The n8n Telegram Pipeline

Every n8n Telegram automation follows the same four-stage shape:

Trigger → Process → Route → Send

Trigger determines when the workflow runs: a Schedule node polling on a fixed interval, a Webhook receiving app events in real time, or an HTTP Request node fetching external data on demand.

Process transforms raw data into something actionable. A Code node (v2, jsCode field) computes scores, calculates percentages, or formats a summary string. When plain templates aren't enough, an OpenAI node generates message bodies tailored to each event's context.

Route is the stage most quick tutorials skip. An IF node or Switch node blocks low-priority records from reaching Telegram. Without this gate, every execution produces a message, the channel fills with noise, and recipients stop reading. The routing stage is what separates a useful alert channel from a muted one.

Send is the Telegram node itself — a single sendMessage call with the formatted text and the destination chat_id.

Step-by-Step: Your First Telegram Alert Workflow

1. Create a Telegram Bot

Open Telegram, search for @BotFather, and send /newbot. Follow the prompts and copy the API token — a string like 7123456789:AAF.... Keep it safe.

To get a chat_id, add the bot to your target group or channel and send any message. Then call https://api.telegram.org/bot<token>/getUpdates in a browser. The chat.id field in the JSON response is your destination. Group chat IDs are negative integers like -1001234567890. This trips people up constantly: passing a positive number routes the message to the wrong chat with no error from the API.

2. Add a Telegram Credential in n8n

In n8n's Credentials panel, create a "Telegram Bot API" credential and paste the token. n8n calls getMe immediately to validate it — you'll see the bot's username if the token is correct.

3. Build the Workflow

Connect your trigger to any processing nodes you need, then place a Telegram node at the end. Set operation to Send Message, set the Chat ID, and set Text to an n8n expression: `={{ $json.alertMessage }}` pulls the formatted string from whichever upstream node computed it.

Run manually first. The execution log shows exactly what text reached the Telegram node. If it looks right in the log, it'll arrive correctly in Telegram.

4. Add the Routing Gate

Place an IF node between your processing nodes and the Telegram node. A condition like `{{ $json.severity }} === "critical"` means only the records that genuinely matter fire an alert. Everything else routes to a second branch that logs to Google Sheets and stops. You don't want volume; you want a channel where every message demands attention.

5. Use Webhook Mode for Incoming Messages

If n8n needs to respond to Telegram messages, use the Telegram Trigger node rather than getUpdates polling. Polling generates a new execution every time n8n checks, even when there's nothing to process. On a self-hosted n8n instance where execution counts matter, that difference adds up fast.

Three Production Patterns

Pattern 1: Scheduled Digest

A Schedule trigger fires on your chosen interval. An HTTP Request node (v4.2) fetches data — support ticket counts, pipeline values, queue depth. A Code node aggregates and formats the output as a single message string. The Telegram node sends it.

Schedule Trigger → HTTP Request (fetch data)
  → Code (format digest string)
  → Telegram (sendMessage)

Useful when volume is the signal, not individual records. A morning push reading "11 tickets open, 2 critical, queue at 84 items" takes five seconds to triage from a phone notification. Checking a dashboard takes three minutes. The difference compounds across a week.

Pattern 2: Real-Time Threshold Alert

A Webhook trigger receives events from your application. A Code node computes a derived value — usage percentage, a risk score, an amount. An IF node gates on whether the threshold is crossed: `{{ $json.usagePct >= 95 }}`. On the true branch, the Telegram node fires an alert with the user name, account tier, and computed value.

Webhook Trigger → Code (compute threshold)
  → IF (>= 95%) ──→ Telegram (fire alert)
                └─→ Google Sheets (log and stop)

The Usage-Triggered Upgrade Engine implements this pattern for SaaS products: it watches for free users approaching their plan ceiling and fires a Telegram alert at 95% usage, giving the sales team a specific name and a timed window to convert before the user churns or upgrades on their own.

Pattern 3: AI-Written Alerts

Same structure as Pattern 2, but an OpenAI node generates the message body instead of a static template. The @n8n/n8n-nodes-langchain.openAi v2.1 node writes context-aware text; the response lives at $json.output[0].content[0].text. For a failing API endpoint, AI names the service and suggests the first debugging step. For a community lead, AI drafts a non-promotional reply tailored to that specific post's context.

Trigger → HTTP Request → Code (evaluate)
  → IF (matches criteria)
  → OpenAI (write message body)
  → Telegram (sendMessage)

The AI Ops Watchtower uses this pattern. It pings up to three configured endpoints every 15 minutes, logs latency history to Google Sheets, and sends an AI-written Telegram alert when something fails. The message isn't "Service X is down" — it's a specific description of what's failing and what to check first, written for the engineer who's going to act on it.

n8n Nodes You'll Use Most

Getting Started with n8n Telegram Templates

Building from scratch takes an afternoon — mostly figuring out chat IDs and expression wiring. A template cuts it to under ten minutes:

1. Browse the full catalog at /templates
2. Find a template matching your use case (three are linked below)
3. Import the workflow JSON into your n8n instance via Settings → Import Workflow
4. Create a Telegram Bot API credential and paste your bot token
5. Update the Config node with your chat_id and threshold values
6. Run the workflow once manually and confirm the Telegram message arrives
7. Activate — the trigger handles scheduling from there

The Community Signal Response Engine is a solid starting point for teams that want Telegram-delivered lead intelligence. It scans configured subreddits every 5 minutes, scores each post 1–10 with AI, and sends a Telegram alert with the post link, score, and a ready-to-post reply draft. All subreddits, keywords, and score thresholds live in a single Config node — no code changes needed.

For ops teams, the AI Ops Watchtower covers endpoint monitoring end-to-end: three configurable URLs, Google Sheets latency history, and AI-written Telegram alerts that describe what's failing rather than just announcing it.

For SaaS products, the Usage-Triggered Upgrade Engine fires Telegram alerts at 95% plan usage. The sales team gets a name and a timing window — the difference between a conversion call and a churned account.

Telegram is the right last step for any alert that can't wait. For patterns that combine OpenAI with live data before the Telegram send, see the n8n AI automation guide. For building the inbound trigger side, n8n webhook automation covers the Webhook node in depth.