softcane

I wanted a valuation tool I could actually audit. Every "AI stock analysis" product I tried either hid the math or hallucinated the inputs. So I built my own.

You type a ticker. You get: - Intrinsic value via DCF using Damodaran's industry datasets (betas, ERP, country risk premiums) - Every assumption exposed — cost of capital, reinvestment rate, terminal value, all of it - LLM-generated bull/bear narratives with cited news sources - The base case and the override case are shown side by side

The math is deterministic. The LLM handles research and narrative Only, it cannot silently change the numbers.

Runs fully local, one Docker command: https://github.com/stockvaluation-io/stockvaluation_io

Rough edges still. Curious what assumptions people would challenge, especially the terminal growth rate, and how to handle high-growth names where DCF tends to break down.