A dark, warm, unhurried room for reading numbers. The craft is borrowed from TradingView and its neighbours — a calm ground, dense data that never shouts, figures that line up to the pixel, thin charts with a soft fill. What is not borrowed is the verdict furniture: no rating gauges, no Strong-Buy dials, no score badges. Instrument craft, not instrument opinions.
Same company, same memo, same numbers as A · The Ledger and B · The Terminal — Northwind Freight at $142.10 — so only the design language differs and the comparison is honest.
Every page carries the same three-column shell and the same agent drawer, with a different specialist selected and a proposal that suits the stage. The funnel on the left navigates the whole set, so you can walk it end to end.
Find candidates. Browse the SEC EDGAR universe. Filters and marks are yours — it never presents a “best” list.
Eliminate quickly. Coarse criteria you rate, and an outcome you select, with a dated reason. Cheap elimination.
Economics, CSI, confidence. The Compound Strength Index — shown only ever beside the components you rated and where the number came from.
BQS and alignment. The second index, and whether it agrees with the first. Compares and questions; never concludes.
Intrinsic value and assumptions. Reverse-DCF as a mirror, your assumptions, and a margin of safety you set yourself.
Dynamic register and triggers. One material risk per line, each with the trigger that would confirm it. User-authored throughout.
Thesis, gate, immutable history. The gate checks whether your own work is complete — not whether the investment is good.
User-owned context and checks. Compares a holding against limits you set. Never suggests a size, weight, rebalance or trade.
Thesis, events, and reviews. Did the trigger you wrote fire? What did you expect, and what have you since observed?
The loop down the left, the current question in the middle, the reverse-DCF and your own readiness on the right. Every answer carries its provenance — sourced, yours, or an AI paraphrase you haven’t checked yet.
The document you wrote before buying. Three figures given room to breathe, the reverse-DCF drawn as a curve with today’s price marked on it, and a bear case you cannot lock the memo without.
My first cut was drawn from the memo-writing loop I designed months ago. The live product is now an investment operating system: a five-stage research funnel, ten named specialist agents in a right-hand drawer, an accept / edit / reject proposal cycle, a screener over the SEC universe, a risk register, allocation context and portfolio monitoring. The redraft is built on that structure instead.
The agent drawer is the part that most changes the design. It is not a chat bolted on — it is where a second opinion is requested, and the rule that keeps the product honest lives in it: validated responses appear here before anything can enter your assessment. So the drawer shows a proposal in a state, with three dispositions, rather than a stream of AI text.
CSI and BQS needed care. A "Compound Strength Index" of 72 looks exactly like the score this product is forbidden to give — until you see that it is arithmetic over five components the owner rated, with the formula pinned and openable. So the index is never drawn alone: the components sit beside it and a line under it says where the number came from. Same data, but the design has to carry the distinction, because the number on its own would read as a verdict.
The look you asked for comes from a category built to produce signals. OwnerMind is built to refuse them — it may show, structure, compare, calculate, journal and question, and must never score, rank, recommend or size. So the borrowing had to be deliberate rather than wholesale.
What survives the constraint is most of the actual craft, because the good part of these tools was never the ratings — it was the calm, the density, and the typography of numbers.
The reverse-DCF is where this is tested hardest, because it is the number most easily mistaken for a target. It is drawn as a mirror: the curve shows what the market is asking you to assume, today’s price is marked on it, and the only thing the app then does is ask “do you believe that?” The answer is a button the owner presses, recorded and dated — not a verdict the product reaches.
Two pages, as asked. Workspace and memo are the two that carry the language furthest — one is dense and working, the other is long-form and finished. Journal and Valuation Lab exist in the other directions and can follow in this one once the language is settled.
Nothing here is a recommendation between directions. A · The Ledger is the one that shipped; this is an alternative, built so it can be judged next to it rather than argued about in the abstract.