Broker tools
IB Risk Navigator · statements
Show execution and raw data, not strategy meaning. Brokers won't build decision support. For them, it's a regulatory liability.
Institutional-grade portfolio intelligence for the investor running a serious book themselves. Every number reconciles to your broker's truth, and you make the call.
A software tool, not investment advice. You control every decision.
If you run a complex book yourself, your tools fail you exactly when it matters: at earnings, on a roll decision, when the call goes in the money.
IB Risk Navigator · statements
Show execution and raw data, not strategy meaning. Brokers won't build decision support. For them, it's a regulatory liability.
spreadsheets · dividend apps
Treat positions as independent rows. They don't understand that a stock plus a short call is one position, not two.
Bloomberg · Aladdin
They solve it, but at €25K–€200K a year with a six-month onboarding. Built for desks, priced out of reach.
institutional-grade, built for one
A decision surface that understands your strategies and reconciles to your broker. Priced for the investor, not the desk.
Other tools model a fair value and hope it's close. Valendis ties every figure to what your broker actually did, and lets you trace it one click deep.
Wired straight into your IBKR gateway. You see current state, not an end-of-day, fifteen-minute-delayed snapshot.
Every figure ties to your actual fills and positions: the broker's truth, not a model's guess.
Coverage, freshness, and the source events are one click away. Honest n/a over faked precision.
It knows your covered call is one position. Multi-leg options read as strategies, not a pile of unrelated rows.
Realized gains, income (dividends, interest, fees), and the mark on what you still hold, each separated and reconciled to your total return. The currency effect is shown, not buried.
Exposure, drift, roll-or-assignment framing, surfaced for the call. You decide; it informs.
AI resets what one tool can do for an investor: classify every instrument in seconds (what once took a reference-data team), read the research behind a name, frame the decision. Classification is where it earns its keep today; the rest is the roadmap. We built the reconciled foundation first, on purpose, so the intelligence always sits on broker truth, not a model's guess.
A flashy demo is a weekend. A system where every number on a real, complex book (preferreds, multi-leg options, currencies, corporate actions) reconciles to your broker and holds up is years of edge-case work. That only comes together at the rare intersection of deep markets-and-risk domain expertise and hands-on agentic-AI engineering. Trust is the slowest thing to copy, and the first thing a serious investor needs.
The tools that could build it can't price it to you without undercutting their own €25K desks. The brokers won't: decision support is a regulatory liability for them. The gap stays open.
You run €500K–€5M across 30–150 positions on Interactive Brokers: equities, bonds, preferreds, covered calls, cash-secured puts. You lose hours a week reconciling, and you don't fully trust your own numbers. Valendis is the cockpit you'd have built, if you had the time.
We hold this as a working hypothesis and we're using this page to test it. If this is you, the one question at the bottom is the most useful thing you can answer.
Valendis is real today: one investor runs it on their own capital, every day. Where it's going is bigger. The reconciled mirror becomes an AI decision cockpit that investigates your book, reads the research and sentiment around every name, and frames the call. You stay the CIO; the AI does the homework. We'd rather show you the seam than paint over it.
Juan Carlos Romero
Fifteen years building market-intelligence, forecasting, and risk decision-support systems in finance and commodities trading, and today leading generative-AI and agentic-AI delivery at a European bank, across 120+ engineers and data scientists. Valendis sits at his intersection: deep markets-and-risk domain knowledge and hands-on agentic-AI engineering. The reconciliation engine is the kind of system he has built before, at scale.
Connect on LinkedIn ▸The fastest way to lose a serious investor's trust is to ask for the keys. Valendis connects to your data with a read-only token, and stays observe-only until you choose to act. Even then, nothing reaches your broker without your explicit approval.
Your data connects read-only. Execution is a separate, opt-in path: off by default, and when you switch it on, Valendis acts only on your explicit, per-order approval. It never trades on its own.
Equities, bonds, preferreds, and multi-leg options, across the whole account. Not just the easy listed names.
Your portfolio truth lives in your own data, reconciled. It isn't locked inside a vendor you can't audit.
Your data is yours. We don't sell it, and we don't train models on it; the AI provider we use doesn't either, by its own API terms. The deeper isolation that keeps your book invisible even to us comes with multi-account access: a commitment we name here before it's built.
Not against Bloomberg. The honest comparison is the spreadsheet and the €30 tracker you use now, plus the hours you spend reconciling instead of deciding.
One account: the reconciled mirror and core analytics. The truth, in one place.
Multi-account, the full decision cockpit, advanced attribution. For the serious book.
Reclaimed from reconciliation. These tiers are a hypothesis we're testing, so tell us where they land.
We're building quietly, with a small group of serious investors. Leave your email and the one answer that helps us most. A person reads every reply.
No spam, no launch noise. We'll only write when there's something real to show.