What We're Building

This guide shows you how to build an AI-powered market analysis and alerting system using Claude Cowork for research and Telegram for delivering alerts. We are not building an automated trading execution system — this is an analysis tool that helps you make more informed decisions.

⚠️ Disclaimer: This is for educational purposes only. Trading involves substantial risk of loss. Never invest money you cannot afford to lose. Past performance does not guarantee future results. Always consult a qualified financial advisor.

System Architecture

The system has three components:

  1. Data Collection — Claude Cowork browses financial sites and extracts market data
  2. Analysis Engine — A Python script processes data and generates insights
  3. Alert Delivery — Telegram Bot API sends analysis summaries to your phone

Step 1: Set Up Telegram Bot

Create a Telegram bot to receive alerts:

  1. Message @BotFather on Telegram
  2. Send /newbot and follow the prompts
  3. Save your bot token (format: 123456:ABCdef...)
  4. Start a chat with your bot and send any message
  5. Get your chat ID via https://api.telegram.org/bot<TOKEN>/getUpdates
# Python: Send a test message
import requests BOT_TOKEN = "your-token"
CHAT_ID = "your-chat-id" def send_alert(message): url = f"https://api.telegram.org/bot{BOT_TOKEN}/sendMessage" requests.post(url, json={ "chat_id": CHAT_ID, "text": message, "parse_mode": "Markdown" })

Step 2: Configure Claude Cowork for Market Research

Set up Claude Cowork to gather market data. Create a workflow prompt:

# Cowork Research Prompt Template:
"Open TradingView in the browser. Navigate to the BTC/USD chart.
Analyze the following:
1. Current price and 24h change
2. RSI, MACD, and Bollinger Band signals
3. Key support and resistance levels
4. Volume trends over the past 7 days Save your analysis as a structured summary in ~/trading/daily-analysis.txt
Include confidence levels (low/medium/high) for each signal."

Cowork will navigate the browser, read the charts visually, and produce a structured analysis file. You can schedule this to run at market open each day.

Step 3: Build the Analysis Engine

import json
from datetime import datetime def analyze_report(report_path): """Parse Cowork's analysis and generate trading signals.""" with open(report_path) as f: analysis = f.read() signals = { "timestamp": datetime.now().isoformat(), "bullish_signals": [], "bearish_signals": [], "neutral_signals": [], "overall_bias": "neutral" } # Simple keyword-based signal extraction bullish_keywords = ["golden cross", "oversold", "support held", "volume surge"] bearish_keywords = ["death cross", "overbought", "resistance rejected", "declining volume"] for keyword in bullish_keywords: if keyword.lower() in analysis.lower(): signals["bullish_signals"].append(keyword) for keyword in bearish_keywords: if keyword.lower() in analysis.lower(): signals["bearish_signals"].append(keyword) # Determine bias bull_count = len(signals["bullish_signals"]) bear_count = len(signals["bearish_signals"]) if bull_count > bear_count + 1: signals["overall_bias"] = "bullish" elif bear_count > bull_count + 1: signals["overall_bias"] = "bearish" return signals

Step 4: Format and Send Alerts

def format_alert(signals): """Format signals into a readable Telegram message.""" emoji = {"bullish": "🟢", "bearish": "🔴", "neutral": "⚪"} message = f"""
📊 *Daily Market Analysis*
🕐 {signals['timestamp'][:16]} {emoji[signals['overall_bias']]} *Overall Bias: {signals['overall_bias'].upper()}* ✅ *Bullish Signals:*
{chr(10).join(f" • {s}" for s in signals['bullish_signals']) or " None detected"} ❌ *Bearish Signals:*
{chr(10).join(f" • {s}" for s in signals['bearish_signals']) or " None detected"} ⚠️ _This is analysis only, not financial advice._
""" return message.strip() # Send the alert
signals = analyze_report("/root/trading/daily-analysis.txt")
alert_text = format_alert(signals)
send_alert(alert_text)

Step 5: Add News Sentiment Analysis

Enhance your analysis by having Cowork check news sources:

# Cowork News Research Prompt:
"Open Google News and search for 'Bitcoin cryptocurrency'.
Read the top 10 headlines. For each:
1. Classify sentiment as positive, negative, or neutral
2. Note the source credibility (major outlet vs blog)
3. Identify any market-moving events Save results to ~/trading/news-sentiment.txt"

Combine chart analysis with news sentiment for a more complete picture. When technical signals and news sentiment align, the confidence level increases significantly.

Step 6: Automate the Pipeline

Use a cron job to run the complete pipeline daily:

# crontab -e
# Run market analysis at 9 AM EST on weekdays
0 9 * * 1-5 /usr/bin/python3 /home/user/trading/run_analysis.py

The run_analysis.py script should: 1) trigger Cowork research, 2) wait for analysis files, 3) run the signal extraction, 4) send Telegram alerts.

Risk Management Rules

  • Paper trade for 30 days minimum before using real money
  • Never risk more than 1-2% of your portfolio on a single trade
  • Set stop-losses on every position — AI analysis can be wrong
  • Don't over-rely on AI signals — use them as one input among many
  • Keep a trading journal — track which AI signals were accurate and which weren't
  • Diversify — don't concentrate in a single asset or strategy

Next Steps

This system gives you structured market analysis delivered to your phone — a significant advantage over manually checking charts. Extend it by:

  • Adding multiple assets (stocks, crypto, forex)
  • Implementing backtesting to evaluate signal accuracy
  • Using OpenClaw for a free, self-hosted alternative to Claude Cowork
  • Building a dashboard with CoderFile.io to visualize historical signal performance

For more AI monetization ideas, check out how to make money with OpenClaw or explore our side project ideas for developers.