Toolinggeneric
telegram-bot
Telegram bot development - chatbots, notifications, AI assistants, and group automation
claude-office-skills/skillsapi.telegram.org
Install
Use with your agent
Install the telegram-bot skill, then use it as build context. Run: npx skills add https://github.com/claude-office-skills/skills --skill telegram-bot. Then read the installed skill.md and follow its guidance to build or refactor my project.
Telegram Bot
Build Telegram bots for chatbots, notifications, AI assistants, and group automation. Based on n8n's Telegram workflow templates.
Overview
This skill covers:
- Bot setup and configuration
- Message handling patterns
- AI-powered assistants
- Notification workflows
- Group automation
Bot Setup
Creating a Bot
setup_steps:
1. create_bot:
- open: @BotFather on Telegram
- command: /newbot
- provide: bot_name
- provide: bot_username (must end in 'bot')
- receive: API_token
2. configure_bot:
- command: /setdescription
- command: /setabouttext
- command: /setuserpic
- command: /setcommands
3. get_chat_id:
- start: conversation with bot
- call: https://api.telegram.org/bot{TOKEN}/getUpdates
- extract: chat.id from response
Bot Commands
commands:
- command: /start
description: "Start the bot"
- command: /help
description: "Show available commands"
- command: /status
description: "Check system status"
- command: /subscribe
description: "Subscribe to notifications"
- command: /unsubscribe
description: "Unsubscribe from notifications"
Message Handlers
Basic Message Handler
workflow: "Telegram Message Handler"
trigger: telegram_message
handlers:
text_message:
action: |
1. Parse message text
2. Determine intent
3. Process request
4. Send response
command:
pattern: "^/"
action: route_to_command_handler
photo:
action: |
1. Download photo
2. Process with vision AI
3. Respond with analysis
document:
action: |
1. Download document
2. Extract content
3. Process and respond
voice:
action: |
1. Download audio
2. Transcribe with Whisper
3. Process text
4. Respond (text or voice)
location:
action: |
1. Extract coordinates
2. Lookup local info
3. Respond with relevant data
n8n Workflow
workflow: "Telegram Bot n8n"
nodes:
- name: "Telegram Trigger"
type: "n8n-nodes-base.telegramTrigger"
parameters:
updates: ["message", "callback_query"]
- name: "Route Message Type"
type: "n8n-nodes-base.switch"
parameters:
rules:
- output: 0
condition: "{{ $json.message.text.startsWith('/') }}"
- output: 1
condition: "{{ $json.message.photo }}"
- output: 2
condition: "{{ $json.message.voice }}"
- output: 3
fallback: true
- name: "Process with AI"
type: "n8n-nodes-base.openAi"
parameters:
model: "gpt-4"
messages:
- role: "system"
content: "You are a helpful Telegram assistant."
- role: "user"
content: "{{ $json.message.text }}"
- name: "Send Response"
type: "n8n-nodes-base.telegram"
parameters:
chatId: "{{ $json.message.chat.id }}"
text: "{{ $json.response }}"
AI-Powered Bot
GPT-4 Integration
ai_bot:
name: "AI Assistant Bot"
system_prompt: |
You are a helpful AI assistant on Telegram.
Guidelines:
- Be concise (Telegram has message limits)
- Use emojis appropriately
- Format with markdown when helpful
- Ask clarifying questions if needed
features:
- conversational_memory: true
- context_window: last_10_messages
- tools: [web_search, calculator, weather]
message_formatting:
max_length: 4096
split_long_messages: true
use_markdown: true
Multi-Modal Bot
multimodal_bot:
handlers:
text:
model: gpt-4
action: chat_completion
image:
model: gpt-4-vision
action: analyze_and_respond
voice:
transcribe: whisper
process: gpt-4
respond: text_or_voice
document:
extract: based_on_type
summarize: gpt-4
respond: text
Notification System
Alert Bot
workflow: "System Alert Bot"
triggers:
- source: monitoring_system
event: alert
- source: ci_cd
event: build_status
- source: ecommerce
event: new_order
notification_templates:
alert:
format: |
๐จ *Alert: {severity}*
*Service:* {service}
*Message:* {message}
*Time:* {timestamp}
[View Dashboard]({dashboard_link})
build:
format: |
{status_emoji} *Build {status}*
*Project:* {project}
*Branch:* {branch}
*Commit:* `{commit_short}`
{details}
order:
format: |
๐ *New Order!*
*Order:* #{order_id}
*Customer:* {customer}
*Total:* ${total}
*Items:* {item_count}
routing:
by_severity:
critical: [admin_group, on_call_user]
warning: [team_group]
info: [logging_channel]
Scheduled Notifications
scheduled_notifications:
daily_digest:
schedule: "9am daily"
template: |
๐ *Daily Summary - {date}*
๐ Sales: ${sales} ({change})
๐ฅ New users: {new_users}
๐ซ Open tickets: {tickets}
Have a great day! โ๏ธ
weekly_report:
schedule: "Monday 9am"
template: weekly_metrics_report
reminder:
trigger: custom_event
template: |
โฐ *Reminder*
{reminder_text}
Scheduled by: {creator}
Group Automation
Welcome Bot
group_bot:
on_member_join:
action: |
1. Check if new member
2. Send welcome message
3. Share rules
4. Suggest introduction
template: |
๐ Welcome to {group_name}, {user_name}!
Please:
1. Read the /rules
2. Introduce yourself
3. Ask questions anytime!
Enjoy your stay! ๐
on_member_leave:
action: optional_goodbye
moderation:
- spam_detection: auto_delete + warn
- link_filtering: whitelist_only
- flood_control: rate_limit
Poll & Survey Bot
poll_bot:
commands:
/poll:
format: "/poll Question | Option1 | Option2 | ..."
action: create_poll
/survey:
action: start_multi_question_survey
/results:
action: show_aggregated_results
features:
- anonymous_voting: optional
- multiple_choice: optional
- scheduled_close: optional
- export_results: to_sheets
Inline Features
Inline Keyboard
inline_keyboard:
example:
text: "Choose an option:"
keyboard:
- row:
- text: "Option 1"
callback_data: "opt_1"
- text: "Option 2"
callback_data: "opt_2"
- row:
- text: "Cancel"
callback_data: "cancel"
callback_handler:
opt_1:
action: process_option_1
response: "You selected Option 1"
opt_2:
action: process_option_2
response: "You selected Option 2"
cancel:
action: cancel_operation
response: "Operation cancelled"
Inline Query
inline_mode:
trigger: "@botname query"
results:
- type: article
title: "{query} result"
description: "Click to send"
message_text: "Result for: {query}"
- type: photo
photo_url: "{search_result_image}"
- type: gif
gif_url: "{gif_url}"
Output Example
Request: "Create a Telegram notification bot for server alerts"
Output:
# Server Alert Telegram Bot
## Setup
1. Create bot with @BotFather
2. Get token: `123456:ABC-DEF...`
3. Get chat ID for alerts channel
## n8n Workflow
```yaml
nodes:
1. Webhook (receives alerts)
2. Format Message
3. Telegram Send
Alert Templates
Critical Alert:
๐จ CRITICAL ALERT
Server: production-web-01
Status: DOWN
Time: 2026-01-30 14:32:05 UTC
Details:
- CPU: 98%
- Memory: 95%
- Disk: 89%
[View Grafana] [Acknowledge]
Recovery:
โ
RECOVERED
Server: production-web-01
Downtime: 5 minutes
Status: All systems normal
Incident resolved automatically.
Implementation
// Send alert function
async function sendAlert(severity, message, details) {
const emoji = {
critical: '๐จ',
warning: 'โ ๏ธ',
info: 'โน๏ธ',
success: 'โ
'
};
const text = `${emoji[severity]} *${severity.toUpperCase()}*\n\n${message}\n\n${details}`;
await telegram.sendMessage({
chat_id: ALERT_CHANNEL_ID,
text: text,
parse_mode: 'Markdown'
});
}
Features
- Severity-based routing
- Inline action buttons
- Acknowledgment tracking
- Escalation rules
---
*Telegram Bot Skill - Part of Claude Office Skills*