wlthy

Allocation breakdown. `by` ∈ {class, sector, geography, currency}. Returns rows with category, value_usd, pct.

get_allocation

Single asset detail + 30-day snapshot history.

get_asset

OLAP-style aggregation. Returns end-of-period values (balance semantics — not summed across days within the period). Dimensions: `net_worth` (account-wide NW per period), `sheets_sections` (groups by the USER'S own sheet names — so Gold under a 'Cash' sheet stays in 'Cash' because that's where the holder put it; use `assets_class` for canonical asset-type grouping instead), `assets_class` (groups by Asset.type: stock / crypto / real_estate / vehicle / precious_metal / private_equity / manual_other / cash / crypto_wallet), `stocks_sector` (sector dimension on stock-class assets only), `stocks_geography` (country dimension on stock-class assets only — use get_allocation(by=geography) for ALL asset classes), `crypto`. grain ∈ {daily, monthly, quarterly}. display ∈ {totals, allocation}.

get_recap

Extract asset positions from a document (PDF / CSV / image). Async fire-and-return: the tool persists the file, kicks off the parse in the background, and returns {job_id, state, confirm_url} immediately. The user opens confirm_url in their browser to review the extracted assets and confirm which to import — this tool NEVER creates assets directly. Base64-only (no URL fetch — closes SSRF surface). Counts against the account's monthly AI-call cap; capped responses come back as isError with error='cap_exceeded'.

parse_document

Paginated asset list. Use `limit` (default 100, max 200) and `offset` for pagination. Filter by `type` or `portfolio_id`. Use `status` to lens the sale state: 'active' (default; live holdings only), 'sold' (history of disposed assets), or 'all' (both — each row exposes `sold_at` so you can badge sold ones).

list_assets

Recurring cash flows (income / expense), item by item: label, kind, category, amount_native, currency, recurrence, start/end date, linked asset/debt id, the DERIVED country (+ country_source), and monthly_equivalent_usd (from the canonical cashflow engine). Filter by `kind` (income/expense), `country`, or `recurrence`. Use this for income/expense analysis, savings rate, debt-service capacity, or cash flow by country.

list_flows

Per-debt list (the debt side of net worth, item by item): balance, currency, balance_usd, the DERIVED country (inherited from the linked asset / explicit override / Unknown) + country_source, the linked asset id+name, and interest_rate_apr / monthly_payment_native / maturity (null when the user hasn't entered them — never guessed). Filter by `type`, `country` (ISO alpha-2 or 'Unknown'), or `linked_asset_id`. Use this for debt-by-country or per-debt detail; get_net_worth only gives the aggregate total.

list_debts

All portfolios in the account. Returns id, name, slug, base_currency.

list_portfolios

Assets, debts AND net worth bucketed by country, with percentages (each group sums to exactly 100.00) + debt-country provenance (explicit / inherited-from-asset / unknown). Debt country is DERIVED from the linked asset (single source of truth); the 'Unknown' bucket stays in the denominator. Prefer this over get_allocation(by=geography) when the user asks for assets AND debts (or net worth) by country.

net_worth_by_country

Total net worth (USD) + 1-day & 30-day delta + asset/debt totals. Pass `portfolio_id` to scope to a single portfolio.

get_net_worth