AI Beneficiary Stock Radar

Python MVP for a daily or weekly AI beneficiary stock early warning report.

It scans configurable public sources, detects AI-related ticker/company mentions, stores evidence in SQLite, scores candidates, and generates an HTML report.

This is a research tool, not a buy/sell recommendation engine.

Time Ledger

The time-ledger/ folder contains a standalone HTML time-reporting app for daily charge-code tracking. Open time-ledger/index.html in a browser to use the app locally.

Setup

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item config.example.yaml config.yaml

Edit config.yaml to change watchlists, tickers, keywords, RSS feeds, and SEC settings.

Run

python -m ai_stock_radar.cli run --config config.yaml

The report will be written to reports/ and a concise console summary will print.

Skip free market-data enrichment if the endpoint is unavailable:

python -m ai_stock_radar.cli run --config config.yaml --skip-market-data

Review the first stored signal for specific names:

python -m ai_stock_radar.cli backtest --config config.yaml --tickers IREN NBIS MRVL AMD

On Windows, you can also use the included runner:

.\scripts\run_radar.ps1

With configured Discord or Telegram webhooks:

.\scripts\run_radar.ps1 -Notify

Notifications

Edit config.yaml:

notifications:
  enabled: true
  discord_webhook_url: "https://discord.com/api/webhooks/..."
  telegram_bot_token: ""
  telegram_chat_id: ""

Leave fields blank for services you do not want to use.

Local Scheduling

The Codex automation already posts reports in this thread. For a local Windows backup, create a Task Scheduler action that runs:

powershell.exe

With arguments:

-ExecutionPolicy Bypass -File "C:\Users\Alan\Documents\Codex\2026-05-11\i-want-to-create-some-sort\scripts\run_radar.ps1"

Current MVP

• Configurable watchlist and candidate ticker list
• RSS/source collection
• SEC company submission collection for configured tickers
• Keyword and ticker detection
• Evidence storage in SQLite
• Basic AI beneficiary scoring
• Price momentum enrichment for early-vs-crowded labeling
• HTML report generation
• First-signal review command for stored evidence
• Optional Discord and Telegram notifications
• PowerShell runner for local scheduling

Planned Next

• Earnings transcript language-shift analysis
• Financial validation
• Email/Telegram/Discord notifications
• X/Twitter monitoring when API access is available
• Backtesting mode against historical winners