get_budget_vs_actual
ChatGPTCompares household budget amounts against actual spending by category and category group for a month, quarter, trailing year, or year-to-date period. The response includes category totals, category-group totals, annual budget context, remaining annual budget, and travelBudgetContext when travel data is available. budgetedToDate is the period-to-date pace target, not the full-year budget. Some categories are intentionally covered by temporary life events (tuition, childcare, etc.) instead of the monthly budget — check each category's lifeEvents and the top-level activeLifeEvents before calling spending unbudgeted.
get_financial_summary
ChatGPTReturns the authenticated household's Per Diem financial snapshot: net worth, account balances, holdings with concentration by security, the current-year annual cash-flow model shown in the Budget page modal, steady-state annual spend (one-time life events excluded), FI year/age from the Monte Carlo FIRE plan with its success rate, and Per Diem's daily discretionary allowance for non-recurring expenses. The currentYearAnnualCashFlow field covers after-tax cash flow, surplus/deficit, taxes, retirement contributions, and draw/contribution schedules. The relatedTools field maps deeper questions to the other Per Diem tools.
get_fire_plan_analysis
ChatGPTGet the household's FIRE plan results from Per Diem's 1,000-run Monte Carlo simulation: FI year/age, per-year retirement success probabilities (success-rate trend), sampled year-by-year ledger projections of income, expenses, net worth, and taxes, the plan's input assumptions (income streams with stop dates, scheduled life events and goals such as college funding or inheritances), plus the simulation methodology and caveats. Use this for questions about retirement confidence, probability of success, or what the projection assumes — not just the single FI year.
get_fire_scenarios
ChatGPTList the household's saved what-if scenarios (FIRE plan forks): what each scenario changes (e.g. quit a job, move states, market drop, spending cut), its last Monte Carlo result (FI year, success rate), and how that compares to the default plan. Pass scenario_id for one scenario's year-by-year projection. Read-only: scenarios are created and recomputed in the Per Diem app, and results marked stale predate the latest plan inputs.
get_net_worth_history
ChatGPTHistorical household net-worth snapshots (daily, plus event-driven captures on meaningful moves) for questions about net-worth trajectory, growth, or volatility over time. Supports daily/weekly/monthly granularity over up to 60 months. Points cover connected accounts only and exclude manually tracked assets like real estate, so the latest point can be lower than get_financial_summary netWorth.
get_spending_history
ChatGPTActual household spending by calendar year for recent years (up to 6): recurring budget-category spend broken down by category group, with tax payments and one-time life-event spending split out. Use this to ground steady-state spending estimates and year-over-year trends in actuals; partial years are flagged.
search_transactions
ChatGPTSearch Per Diem transactions by date range, merchant/name query, optional category name, and optional category_group such as TRAVEL. Defaults to the most recent 30 days. Returned transactions include signedAmount and cashFlowDirection; use those to distinguish spending outflows from income/transfer inflows.
