For Professionals • Financial Planning
A household plan grounded in your facts — not someone else’s training data.
Net worth, retirement projection, scenario lab, fiduciary-priority recommendations. A deterministic engine produces every number. The AI explains them — never invents them. PII is tokenized before any frontier model sees it.
Net worth and cash-flow picture
Add household members, income, expenses, assets, and liabilities. The engine computes net worth, monthly surplus, savings rate, and emergency-fund coverage on every change.
Retirement projection
Years to retirement, projected portfolio at retirement age, safe annual income vs. target, and a clear readiness band: on track, moderate, off track, or unknown.
Scenario lab
Seven levers (retire age, monthly savings, spending, salary growth, returns, home purchase, income shock). Compare baseline-vs-scenario delta with a recommendation diff.
Fiduciary recommendations
Ordered by what actually matters: solvency → downside protection → match capture → debt → compounding → optimization. Every recommendation cites the facts and assumptions it relied on.
Why this isn’t another budgeting app
- LLM synthesizes; engine calculates. A pure TypeScript financial engine is the only source of numbers. The AI explains them — never invents them. No hallucinated retirement projections.
- PII tokenization at the boundary. Names, addresses, employer names, account labels are replaced with
[[PERSON_1]]-style tokens before any call to Claude / GPT / Gemini. Re-hydrated on the way back. - Frozen-snapshot reports. PDF exports regenerate byte-for-byte from a typed input tuple. Forward to your CPA; "why does this say X?" has a real answer.
- Honest uncertainty. When data’s missing, recommendations say so. When a calculation uses a proxy (cash as a stand-in for "savings earmarked toward this goal"), every fact list and caveat sentence labels it.
Pairs with PrivDF’s PDF tools
Need to add an account from a brokerage statement? The planner sends you to PrivDF, you draw a rectangle around the holdings table, the CSV comes back — pre-filled into your account form. The statement bytes never leave your browser at any point.
In the planner: click Upload statement on the Accounts page.
PrivDF opens with the statement-extract preset. Draw a rectangle around the holdings table.
Click Send to planner. The CSV pre-fills your account form. You confirm and save.
CSV passes between the two apps via postMessage — nothing touches the network.
Privacy posture
- No frontier-model PII leakage. Names, addresses, account labels, employer names, dependent names are tokenized before any call to Claude, GPT, or Gemini. The token map exists only for the duration of one request.
- Audit trail. Every household mutation, every AI response, every report event writes an audit record — with the anonymized payload, never raw PII.
- Hard delete. Privacy center exposes a one-click "delete everything" flow plus a JSON export of all data we hold for your household.
Ready to plan?
No account aggregation. No Plaid. Manual entry, fiduciary-grade output. Free during early access.
Open the planner →