Skip to main content

​
Overview

The Telegram bot provides a conversational interface for analyzing business URLs. It includes built-in rate limiting, status monitoring, and formatted output optimized for mobile devices.

​
Starting the Bot

​
Prerequisites

  1. Create a Telegram bot via @BotFather
  2. Add your bot token to .env:
TELEGRAM_BOT_TOKEN=your_telegram_bot_token_here
  1. Ensure all other required environment variables are configured (see Configuration)

​
Launch Command

Start the bot server: 
python telegram_bot.py
Expected output: 
Lead Intelligence Engine -- Application running...
The bot runs continuously. Keep the terminal open or use a process manager like pm2 or systemd for production deployments.

​
Running in Background

nohup python telegram_bot.py > bot.log 2>&1 &
Stop the bot:
pkill -f telegram_bot.py

​
Bot Commands

​
/start - Welcome Message

Displays bot capabilities and available commands: 
Welcome! I'm the Lead Intelligence Engine Bot. Generate leads for reachout purposes

Commands:
/analyze <url> - Analyze a business website
/model - Show the current LLM model
/status - Show the current status of the AI

​
/analyze - Analyze a Business URL

Analyze a business website and add qualified leads to your CRM. Syntax: 
/analyze https://example.com
Example Conversation:
1

Send Command

/analyze https://austinplumbing.com
2

Bot Responds with Processing Status

Analyzing https://austinplumbing.com...
This may take up to 20 seconds.
3

Bot Returns Analysis Result

Analysis Complete for Austin Plumbing Solutions

Type: Plumbing Service
Primary Service: Foundation Package
Fit Score: 85/100

Reasoning: Local service business with weak online presence. Foundation Package would establish professional web identity.

Outreach Angle: Help local customers find you online with a mobile-responsive website that showcases your services and builds trust.

**Latency:** 12.34s

Result saved to Coda CRM.
You can also send URLs directly without the /analyze command. The bot automatically detects and processes any message starting with http:// or https://.

​
/model - Show Current LLM Model

Displays the language model being used for analysis: 
Current LLM Model:
`llama-3.3-70b-versatile`
This helps verify which Groq model is active. The default is llama-3.3-70b-versatile.

​
/status - AI System Status

Shows comprehensive system status including token usage and rate limits: 
SYSTEM STATUS REPORT
━━━━━━━━━━━━━━━━━━━━
STATUS: System Online [ONLINE]
LLM QUOTA: WITHIN LIMITS
MODEL: `llama-3.3-70b-versatile`
━━━━━━━━━━━━━━━━━━━━

TOKEN CONSUMPTION
• Total: `45,203`
• Prompt: `32,145`
• Completion: `13,058`

LAST ACTIVITY: `2026-03-11 14:32:45`
━━━━━━━━━━━━━━━━━━━━
_Updates are tracked per session._
Token consumption resets when the bot restarts. This tracks usage for the current session only.

​
Rate Limiting

The bot implements user-level rate limiting to prevent abuse:
  • Limit: 3 requests per minute per user
  • Scope: Per user ID (multiple users can analyze simultaneously)
  • Reset: Rolling 60-second window

​
Rate Limit Response

When exceeded: 
RATE LIMIT EXCEEDED. Please wait a minute before analyzing more URLs.
Rate limits are enforced per user. If you need higher limits for production use, modify MAX_REQUESTS_PER_PERIOD in telegram_bot.py:23.

​
Customizing Rate Limits

Edit telegram_bot.py:
telegram_bot.py
RATE_LIMIT_MINUTES = 1
MAX_REQUESTS_PER_PERIOD = 3  # Change this value
Example configurations:
  • Stricter: MAX_REQUESTS_PER_PERIOD = 1 (1 per minute)
  • Relaxed: MAX_REQUESTS_PER_PERIOD = 10 (10 per minute)
  • No limit: Comment out rate limit check in process_lead_analysis() (not recommended)

​
Example Conversations

​
Successful Analysis

User: https://localcafe.com

Bot: Analyzing https://localcafe.com...
     This may take up to 20 seconds.
     
     [Updates to:]
     
     Analysis Complete for The Local Cafe
     
     Type: Food & Beverage - Cafe
     Primary Service: Foundation Package
     Fit Score: 78/100
     
     Reasoning: Small F&B business with minimal web presence. A simple website would help attract local customers and showcase menu items.
     
     Outreach Angle: Increase foot traffic with an online menu and location map that makes it easy for customers to find you.
     
     **Latency:** 11.2s
     
     Result saved to Coda CRM.

​
Duplicate Detection

User: /analyze https://already-exists.com

Bot: Analyzing https://already-exists.com...
     This may take up to 20 seconds.
     
     [Updates to:]
     
     URL already exists in CRM:
     https://already-exists.com
     
     Duplicate found in CRM

User: https://facebook.com/localbusiness

Bot: Analyzing https://facebook.com/localbusiness...
     This may take up to 20 seconds.
     
     [Updates to:]
     
     Error processing https://facebook.com/localbusiness:
     
     Facebook public metadata unavailable. Please provide website link if available.
Facebook pages often block automated scraping. Ask for the business’s actual website if available.

​
Error Handling

The bot gracefully handles various error scenarios:

​
Missing URL Argument

User: /analyze

Bot: Please provide a URL. Usage: /analyze <url>

​
Invalid URL Format

User: not-a-url

Bot: I only analyze URLs. Use /analyze <url> or just send the link.

​
Network/Extraction Errors

Error processing https://example.com:

No content could be extracted from the URL.

​
Rate Limit Errors (AI Quota)

Error processing https://example.com:

AI Service (Groq) Error: rate_limit_exceeded
Check /status to confirm quota status: 
STATUS: Rate Limited / Quota Reached [OFFLINE/LIMITED]
LLM QUOTA: RATE LIMIT REACHED / OUT OF TOKENS

​
Monitoring & Logs

The bot uses Python’s logging system configured to WARNING level to reduce noise.

​
Viewing Logs

python telegram_bot.py
Logs print to terminal in real-time.

​
Log Levels

To increase logging verbosity, edit telegram_bot.py:14: 
logging.basicConfig(
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
    level=logging.INFO  # Change from WARNING to INFO or DEBUG
)

​
Troubleshooting

​
”CRITICAL: TELEGRAM_BOT_TOKEN not found in .env”

Your .env file is missing the bot token: 
TELEGRAM_BOT_TOKEN=1234567890:ABCdefGHIjklMNOpqrsTUVwxyz
Get a token from @BotFather.

​
Bot Doesn’t Respond

  1. Verify the bot is running: check terminal/logs for “Application running…”
  2. Check bot token is valid: test with /start command
  3. Ensure bot isn’t rate limited on Telegram’s side (restart bot if needed)
  4. Review logs for Python exceptions

​
Analysis Takes >20 Seconds

Same as CLI troubleshooting:
  1. Check network connectivity
  2. Verify target site loads quickly
  3. Monitor Groq API status
  4. Consider websites with less content

​
”INTERNAL ERROR: An error occurred”

Generic error handler triggered. Check logs for full stack trace: 
tail -f bot.log | grep ERROR

​
Production Deployment

Use Process Manager

Deploy with PM2, systemd, or Docker to ensure the bot auto-restarts on failure

Monitor Resources

Track memory and CPU usage. The bot is lightweight but long-running processes can leak memory

Set Up Alerts

Configure alerts for bot downtime or error spikes

Regular Restarts

Schedule daily restarts to clear session state and prevent memory bloat

​
Docker Deployment

Dockerfile
FROM python:3.11-slim

WORKDIR /app

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY . .

CMD ["python", "telegram_bot.py"]
Build and run: 
docker build -t lead-bot .
docker run -d --name lead-bot --env-file .env lead-bot

​
Next Steps

Configuration

Customize bot behavior with environment variables

Coda Integration

Set up your CRM to receive bot-analyzed leads

Build docs developers (and LLMs) love

Get started for freeTalk to us