Automatic financial anomaly detection, earnings alerts, cash flow monitoring. Deploy in 3 steps. No coding required.
Who I am: I'm a growth stock value investor. My nightmare: missing a key financial anomaly — supply chain disruption, gross margin compression, skyrocketing accounts receivable — these are buy warnings I can't afford to miss.
Problem: I can't monitor 1,000 companies 24/7. Solution: I deployed a Python script on a cheap VPS. Every market close, it scans my watchlist, detects anomalies, and pushes alerts to my phone.
Today I'm open-sourcing this. You don't need to know Python. Just 3 commands. You'll have your own investment sentinel.
Why You Need This (And Why You Probably Already Tried Everything Else)
You've probably tried:
Setting up custom alerts on your broker's platform → clunky, limited filtering
Paying for a financial data service → $50-500/month, black box logic
Writing Python scripts yourself → got halfway, gave up on environment setup
What you really need: A custom system that runs 24/7, scans YOUR watchlist, sends alerts to YOUR phone, and you own the entire logic.
✅ Maintains a beautiful web dashboard you can check anytime
✅ 100% transparent — you see exactly which metrics triggered each alert
✅ Runs entirely on your private VPS — zero third-party access to your watchlist
Complete Setup Guide (3 Steps, 5 Minutes)
1
Step 1: Rent a Tiny VPS ($3/month)
Why VPS instead of your laptop? Your laptop sleeps. Your VPS runs 24/7. The sentinel needs to work every single market close, whether you're awake or not.
Disclaimer: I get a small affiliate commission if you use my link. But I've genuinely used Bandwagon for 5 years — this isn't a sellout, it's a honest recommendation. The VPS will be *yours* — I never touch it.
💰
Total cost: $3-6 per month. That's less than a coffee. But it runs 730+ hours per month scanning your portfolio.
2
Step 2: Copy-Paste 3 Commands (No Coding)
Once you have your VPS, login via SSH:
ssh root@your-vps-ip-address
Then run these three commands:
# Command 1: Clone the repository
git clone https://github.com/your-username/investment-sentinel.git
cd investment-sentinel
# Command 2: Configure your watchlist (edit the simple JSON file)
nano config/watchlist.json
# Command 3: Deploy with Docker
docker compose up -d
What happens:
Command 1 downloads the entire sentinel system
Command 2 opens a simple text editor. Paste your stock tickers (one per line). Save and exit.
Command 3 launches everything. Docker handles Python, dependencies, database — you never touch any of it.
That's it. The sentinel is now running 24/7.
⚡
If you can copy-paste and edit a text file, you can do this. Seriously. No Python, no environment variables, no package managers.
3
Step 3: Open Your Dashboard (1 Minute)
In your browser, go to:
http://your-vps-ip-address:8080
You'll see a beautiful dashboard showing:
📊 All your watchlist companies with latest financials
🚨 Anomalies detected (highlighted in red)
📈 Trend charts for key metrics (margins, cash flow, debt)
⏰ Last update time (so you know data freshness)
What to do: Bookmark this page. Check it during your morning coffee, or set up push notifications:
Telegram alerts (free, instant, on your phone)
Email digests (daily summary of all anomalies)
Slack webhook (if you're in Slack all day anyway)
The sentinel continues running in the background. Every market close, it updates.
🎯
From this moment on, you'll never miss a financial anomaly again. Because the system never sleeps.
What Financial Anomalies Does It Detect?
The sentinel monitors 20+ metrics and flags when they deviate from the company's historical patterns:
Debt spike: Total debt increased >20% in one quarter
Inventory buildup: Inventory growing faster than COGS (demand weakening?)
ROIC deterioration: Return on invested capital declined >300 bps
...and 14 more metrics
Each alert shows: what changed, by how much, why it matters. So you're not just getting a red flag — you're getting *context*.
Walkthrough: Real Example
Let's say you're watching Montage Technology (a Chinese semiconductor company you're considering buying).
Day 1: You add 蒙塔基 to your watchlist via the config file. Sentinel starts monitoring.
Day 7: Montage reports Q3 earnings. Sentinel fetches the filings automatically (you don't touch anything).
That evening: You get a Telegram alert:
🚨 MONTAGE TECHNOLOGY: Gross Margin Compression
Q3 2026: 42.3% → Q3 2025: 47.8% (-540 bps)
Why it matters: Potential pricing power loss or input cost inflation
Dashboard: [link]
You open your dashboard. You see Montage's full Q3 picture. Margins down, but cash flow still strong. Probably just temporary pricing pressure from a price war. Or you dig deeper and discover it's an inventory write-down — completely different risk profile.
Without the sentinel? You'd read the earnings call, miss the gross margin trend, buy the stock thinking "all is well," and get surprised a quarter later when the margin pressure persists. By then you're down 15%.
With the sentinel? You caught it on day one. You had a week to decide if it's a buying opportunity or a value trap.
Common Questions
Q: What if I travel or my laptop dies?
The sentinel runs on the VPS, not your laptop. It doesn't care. You can access the dashboard from any device with internet. Alerts go to your phone regardless.
Q: Is this legal? Can I get data-banned?
100% legal. You're fetching publicly available filings from SEC/HK Stock Exchange/Shanghai Stock Exchange. It's exactly what Bloomberg terminals and financial websites do. No API terms violated.
Q: What if I want to add new companies mid-quarter?
Edit the config file, save. The sentinel picks it up within 5 minutes. Add or remove companies anytime — no restart needed.
Q: I'm not technical at all. Can I really do this?
If you can:
Copy and paste text
SSH into a server (we'll give you the exact command)
Edit a JSON file in a text editor (we'll show you exactly what to change)
...then yes. Seriously. This is the whole point — we designed it for people who are smart investors but not software engineers.
⚠️ One real limitation: The sentinel works best for companies that file quarterly reports (US SEC, HK Stock Exchange). If you're watching small-cap stocks that barely report, you won't have data to analyze.