We didn't start
with the answer.

Higher timeframes stalled. Cache hydration raced the interface. Charts could show stale windows. Instead of hiding those cracks, we worked through them: cache-first reads, shared Redis state, safer backfills, clearer freshness, tighter security and hundreds of small loading-path repairs.

Then we tried two cleaner architectures. One became a production snapshot; another became a modular FastAPI and HTMX rewrite. Neither was treated as destiny. The useful ideas survived and the product moved forward in the current repository.

The dashboard was realigned around the validated stack: Flow Board, Market State, Rotation, the market-neutral book and a Track Record that separates a blue backtest claim from green live proof. Pricing and onboarding followed the product that actually existed — not the one we had imagined months earlier.

Aegium went public on 19 June. Since then, the work has continued in the open: shadow engines compete on common windows, the liquid universe maintains itself, rejected research stays rejected, and execution fidelity is measured across venues instead of assumed.

The story is no longer about how much was built. It is about how much survived contact with data, markets and time.

A paper record has one weakness: the publisher can rewrite it. Since September, every frozen daily brief is hashed, chained to the day before it, and timestamped into the Bitcoin blockchain. The chain, the exact bytes and the proofs are public on the data page, so anyone can check that a day’s figures are the ones published that day. The Telegram channel stopped predicting and now posts only what the ledger records.

The same month, a guard added to every equity writer found a run of August marks in which one book had briefly fallen out of the valuation. The flagship’s lifetime drawdown had been overstated for a month. The marks were removed after a backup, the return was unchanged, and the guard now refuses such a mark before it is written. June’s correction cost us an attractive number; this one gave one back. A process that only runs in one direction is not a process.

Less product, more record. Nothing is sold while the record builds, and every account sees every chart. The market tools left the product and became internal instruments; what a signed-in account watches is one engine, shown as outcomes: equity, drawdown and trades, with the method kept in-house. Research studies are now pre-registered and hashed before they run, so a result cannot quietly redefine the question it was asked.