TFP - Timmy's Financial Planner

TFP is a command-line Python app that reads a plan.json financial plan and generates a self-contained report.html.

This workflow is designed for collaborating with an AI agent:

1. You and the AI agent edit plan.json.
2. TFP regenerates report.html.
3. You refresh the browser to see updates.

Quick Start

1. Ensure Python 3.10+ is installed.
2. From the repo root, run:

python3 -m tfp plan.json --server

By default this:

• Writes/updates report.html
• Serves it at http://127.0.0.1:8000/report.html
• Watches plan.json for changes and regenerates the report automatically

Working With an AI Agent

Use this loop:

1. Ask the AI agent to update plan.json (accounts, income, expenses, taxes, assumptions, etc.).
2. Save the file.
3. TFP detects the change and rebuilds the report.
4. Refresh your browser to review the updated charts/tables.
5. Repeat until the plan looks right.

Example prompts for the AI agent:

• "Increase annual salary to 180000 and extend retirement date to 2040-06."
• "Add a Roth conversion of 40000/year from 2030-01 to 2037-12."
• "Lower inflation to 2.5% and rerun assumptions."

Recommended Command Options

python3 -m tfp plan.json --server --mode deterministic --port 8000 --watch-interval 0.5

Useful flags:

• -o, --output: output HTML filename/path (default: report.html)
• --host: server bind host (default: 127.0.0.1)
• --port: server port (default: 8000)
• --watch-interval: polling interval in seconds (default: 1.0)
• --mode: override simulation mode (deterministic, monte_carlo, historical)

Validation and One-Off Runs

Validate only:

python3 -m tfp plan.json --validate

Generate once without server mode:

python3 -m tfp plan.json -o report.html

Troubleshooting

• If startup fails with validation errors, fix the reported JSON paths in plan.json and rerun.
• If --server is running but the browser is stale, refresh the page.
• If the port is in use, pick another one (for example --port 8001).