DEV Community

Văn Tuấn Lê
Văn Tuấn Lê

Posted on

I built a free AI investing toolkit after running 7 experiments with 8 AI models — here's what I learned

I've been running an experiment for months: asking 8 different AI models (GPT-4, Claude, Gemini, DeepSeek, Kimi, Grok, Mistral, Perplexity) the exact same investing questions.

One pattern kept emerging: every model, regardless of architecture, fell into one of 4 investing philosophies:

  • Index school — diversify, minimize fees, let the market work
  • Value school — find underpriced assets with a margin of safety
  • Growth school — bet on high-growth compounders
  • Trend school — follow price momentum and manage risk tightly

This got me thinking: most people using AI for investing are prompting blindly — their prompts don't match their actual philosophy. So they get generic answers.

What I built

A free browser-based toolkit: My AI Investment Operating System

Main tool: 5-question diagnostic that identifies your investing school and generates a personalized Prompt Library — prompts you can drop into Claude, GPT-4, or any AI today.

6 companion tools:

  • DCA Simulator — model dollar-cost averaging scenarios with compound projections
  • Recovery Navigator — calculate breakeven points and map exit strategies for losing positions
  • Dividend Engine — screen dividend stocks through the lens of your investing school
  • Kelly Master — apply Kelly Criterion to size positions correctly
  • Pyramid Builder — plan staged entry positions at multiple price levels
  • Portfolio Clarity — audit your current holdings against your school's criteria

Tech stack

Pure HTML/CSS/JS, no framework, no backend. Everything runs client-side. Hosted as static files on a WordPress site's /tools/ directory.

The AI committee logic uses a scoring matrix built from patterns across all 7 experiments — no live API calls, all logic is embedded.

Link

All free, no login: https://ordinarymantrying.com/tools/ai-invest-os.html

Would love feedback from developers who also invest — especially on the UX of the diagnostic flow.

Top comments (0)