driftcurve

What driftcurve does

driftcurve models household retirement and financial-independence scenarios. A household profile may contain multiple people, individual retirement timing and savings rates, plus account categories, liabilities, mortgages, pensions, Social Security and property-sale details.

It projects income, savings growth and withdrawals, with basic tax modeling and scenario comparisons. Its Monte Carlo runs produce forecasts, while the guardrail solver works out monthly spending against portfolio thresholds and success odds.

Fees and plans

The available plan is Free. Local use is self-managed, carries no subscription charge and does not require an account. No trial plan is published.

You must handle the installation yourself, including Python 3.10 or later and the listed dependencies.

Where it runs

You can use driftcurve through a web application, a self-hosted deployment or its API. The quick start creates a Python virtual environment, installs requirements and starts Uvicorn at localhost:8000.

The documented API includes forecast and guardrails routes, together with authentication, profile and scenario routes for hosted multi-user mode. No native iPhone or Android app is specified, and the facts do not describe phone-specific functions.

Protection, as published

The maker presents driftcurve as a modeling and educational tool rather than a licensed financial, investment, tax or legal adviser. Its code is available under Apache License 2.0.

Under the standard local configuration, browser localStorage holds household information and saved scenarios; forecasts run through the Python backend, without an account or server-side database. Optional hosted mode stores profiles and scenarios in driftcurve.db, uses email/password accounts and server-side sessions, and hashes passwords with PBKDF2-SHA256. The provider does not publish claims here about regulators, deposit protection, two-factor sign-in or dispute handling.

Strengths

The model covers a whole household, including separate retirement dates and savings rates, many asset types, debts, pensions, Social Security and property sales.

Historical simulations resample annual data from 1966 through 2024 across equities, Treasuries, residential real estate, inflation and Bitcoin. You can recenter returns to long-run targets while retaining historical volatility and cross-asset correlations, or use static annual rates.

The Compare page shows two scenarios together on a single chart. Guardrails target an 0.80 probability of success by default, with documented upper and lower trigger probabilities of 0.99 and 0.25.

Catches and limits

Guardrail calculations require a retirement year and a positive liquid portfolio. The README estimates one such computation at about 20–45 seconds.

Default forecasts are projections from historical data and your assumptions, not predictions or guarantees. Account synchronization is not provided.

Local use requires installation and a Python environment. Hosted mode changes the data arrangement by enabling accounts, sessions and database-backed profiles and scenarios.

Who may want another tool

People seeking a hosted consumer service, automatic account connections or a native phone app may not find those functions here.

Verdict

Technically minded households can run driftcurve locally or expose it through an API to build no-subscription retirement models. Its household inputs, historical resampling, return controls and spending guardrails support detailed scenario work.

It is less suitable for someone who expects connected financial accounts, a ready-made mobile app, or a provider offering regulated advice. The hosted option also introduces stored profiles and scenarios, unlike the default browser-only mode.

Questions