A memory built around you.
One grounded layer at a time.
The product direction is ambitious. This page separates what works today from what needs to be connected and validated before a paid launch.
Replay & reflect
- Brand, landing page, and responsive workspace
- Closed-position CSV imports with validation and duplicate detection
- Replay on stored one-minute bars for the symbols the dataset covers, or bars you upload
- A window for when your fill happened, when the statement kept only the date
- A chat grounded in your own record and held to the replay cursor
- Behavioural findings, each with the control it survived and what it cannot tell
- A profile wheel comparing where you place yourself with what the record shows
- An indicator library, including formulas you write, that never looks past the cursor
- Saved reflections, journal, and export
- Editable identity, markets, specialty, strengths, blind spots, goals, setups, and risk rules
- Descriptive insights linked to trades
- Browser dictation where supported
Context that compounds
- Option price history, so an option replays on its own price rather than its underlying
- Broker-native fills, partial exits, and corrections
- Suggested memories with confirm, reject, and forget controls
- Weekly reviews and goal progress
- Scenario backtesting: alternative entry and exit rules with fees, slippage, data provenance, and out-of-sample validation
Ready for customers
- Public customer authentication and account recovery
- Subscriptions, usage limits, cancellation, and receipts
- Market-data redistribution agreements
- Final privacy policy, terms, and support contact
- Monitoring, backups, recovery, and tenant-isolation tests
- Measured AI and data cost per subscriber
How personalization should work
Every piece of knowledge has a source. Executions establish what happened. The trader's words establish what they say they intended. Statistical observations remain tentative and never silently become personal facts.
- Confirmed profile: self-reported identity, markets, specialty, strengths, blind spots, goals, setups, risk rules, and review preferences.
- Observed behavior: computed sizing, holding periods, exits, and fees linked to trades.
- Suggested interpretation: a question the trader can accept, correct, or reject.
- Changing context: version goals over time; preserve prior plans.
A careful learning loop
Import, normalize, reconstruct, review together, confirm context, calculate observations, then test the next change on new trades. A good review helps inspect decisions; it does not label personality or promise future returns.
The first paid plan
$9.99/month is the founding price. Final allowances depend on data licensing, inference costs, and actual review usage. There is no unlimited-AI promise.
Download the product blueprint