Be Rich or Avoid Bankruptcy
StockManager is a self-hosted Progressive Web App (PWA) designed to help individuals manage their stock portfolios.
Smartly, privately, and without reliance on cloud-based services or third-party APIs.
A lightweight, local-first application to
- track
- analyze
- support decision-making
For stock portfolios.
Using simple interpretable machine learning techniques.
Built to run entirely on your own machine.
No data sent to the cloud, and no bloatware dependencies.
A smart portfolio tracker does more than display charts — it:
-
Analyzes your historical and current holdings
-
Detects risk patterns and price movement trends
-
Applies lightweight ML models (like decision trees) to propose actions
-
Offers offline-first predictive capability
-
Respects user control instead the them own decision
We do not use OpenAI, Claude, Gemini or other API-heavy "AI" solutions.
Instead, we focus on transparent, CPU-only, and locally executed intelligence.
- Decision Tree for rule-based actions (e.g. BUY/SELL/HOLD)
- Simple Predictive Models (linear regression, volatility scoring)
- Portfolio Health Check (diversification ratio, exposure limits)
- Black-box neural networks
- GPU-hungry models
- API-based cloud inference engines
-
Stock Discovery module
-
Support for non-stock-based investments
-
Connect to banks via OpenFinance API
-
Connect to brokers via their Web APIs
-
Manual import via CSV / JSON
The core of StockManager is automated portfolio management through explicit rules defined by the user.
- Stop Loss — sell when price drops below a defined threshold
- Take Profit (Stop Gain) — sell when price reaches a desired target
- Trailing Stop — dynamically sell when price drops X% from peak
- Date-based Rules — sell if no appreciation by a specific date
- Conditional Buy/Hold/Sell — based on model predictions or thresholds
- Rules are defined and prioritized by the user
- Actions may be automatically executed by the system, according to configuration
- All automated logic is explainable and deterministic
- The system logs and justifies all actions for auditability
- Automatic actions are only performed if rules are enabled and configured
- Each rule must be explicitly authorized by the user
- The system is non-opinionated: it does not assume — it executes what you've defined
The StockManager Discover module identifies a stock matching your criteria:
Stock: AAPL34 P/E Ratio: 10.5 Dividend Yield: 4.5% Current Price: R$38.00
50-day Moving Average: R$36.50
The system evaluates your current cash balance and portfolio exposure:
Available cash: R$5,000
Maximum allocation per stock (user-defined): 20% of portfolio
Current exposure to AAPL34: 0%
Because the conditions meet your rules and there is sufficient cash, the system proceeds to:
Automatically place a BUY order for AAPL34 at the target entry price of R$37.00
Create associated Stop Loss (5%) and Take Profit (15%) rules
Log the purchase and rule creation with timestamps and decision metadata
The system notifies you:
"Order placed: BUY 135 shares of AAPL34 at R$37.00 Stop Loss set at R$35.15, Take Profit at R$42.55. This action was performed automatically based on your configured rules and available balance."
You bought PETR4 at R$32.50 Current market price is R$30.75
You configured the following rules:
Stop Loss at R$30.00
Take Profit at R$36.00
Trailing Stop at 5%
The system:
Monitors the asset in real-time (offline/local)
Detects that price is nearing Stop Loss
Triggers a SELL action automatically (or notifies you, depending on your rule priority)
Logs the reason and logic used (e.g. "Trailing Stop triggered: price dropped 5.1% from peak")
✅ All decisions are traceable, local, and under your control
- 100% self-hosted — no external servers
- Local-only processing — no telemetry or background sync
- No vendor dependency — everything runs on your machine
- Your rules, your data, your decisions