driftcurve
github.com
- Free plan
- Open source
- Web
- API
- Self-hosted
What driftcurve is
driftcurve is a free, open-source retirement planning engine for modeling household finances and comparing retirement scenarios. Its inputs can include multiple household members, accounts, debts, mortgages, pensions, Social Security, and property sales. By default, simulations resample annual market data from 1966–2024 across equities, Treasuries, residential real estate, inflation, and Bitcoin. Users can adjust historical returns toward their own long-run targets or enter static annual assumptions. Risk-based spending guardrails calculate monthly spending and suggest changes based on portfolio thresholds and the probability of success. A Compare page displays two scenarios on one chart. In default local mode, household data and saved scenarios stay in browser localStorage, while forecasts run through a Python backend; there is no account or server-side database persistence in that mode. Optional multi-user mode adds email/password accounts, server-side sessions, and profiles and scenarios in driftcurve.db. Setup uses a Python virtual environment, requirements installation, and Uvicorn, then opens at localhost:8000. The project is licensed under Apache 2.0 and is described as an educational modeling tool, not financial, investment, tax, or legal advice.
Who it suits
driftcurve is for households exploring retirement and financial-independence scenarios, including people who want to model multiple household members and income sources. It may also suit users comfortable setting up a self-hosted application.
Good to know, and what to watch
- Models multiple household members and account types.
- Includes Social Security and pension inputs.
- Guardrails connect spending to portfolio thresholds.
- Compare displays two scenarios on one chart.
- Open source under Apache License 2.0.
- Default local mode has no account synchronization.
- Guardrails require a retirement year and positive liquid portfolio.
- Guardrail computation takes about 20–45 seconds.
- The maker says it is not financial advice.
Plans
Every plan and priceFree
Free plan
The verdict
driftcurve offers detailed household retirement modeling, including scenario comparisons and spending guardrails. It is an educational modeling tool, not a substitute for licensed financial, tax, investment, or legal advice.
What the maker publishes
- Account coverage
- Inputs include traditional, Roth, taxable, cash, real-estate and cryptocurrency assets. source
- Apache permissions
- Apache 2.0 grants perpetual, worldwide, royalty-free rights to reproduce, modify, display, perform, sublicense and distribute the work subject to its conditions. source
- API
- The documented API includes forecast and guardrails endpoints, plus authentication, profile, and scenario endpoints for multi-user mode. source
- Authentication security
- The multi-user implementation hashes passwords with PBKDF2-SHA256 and uses server-side session cookies. source
- Calculation limit
- Guardrails require a retirement year and positive liquid portfolio, and the README says a computation takes about 20–45 seconds. source
- Debt and income inputs
- The scenario model includes mortgages, HELOCs, student loans, pensions, Social Security and property sales. source
- Dependencies
- The project declares FastAPI, Uvicorn, aiofiles and SQLAlchemy as runtime dependencies. source
- Developer API
- The repository documents /config, /forecast, /guardrails, authentication, profile and scenario endpoints. source
- Forecast caveat
- The maker says forecasts are projections based on historical data and user assumptions, not predictions, guarantees or professional financial advice. source
- Guardrail API
- POST /guardrails accepts the same scenario body as /forecast and returns spending, thresholds, achieved probabilities and configuration. source
- Guardrail defaults
- The documented defaults target 0.80 probability of success, with upper and lower trigger probabilities of 0.99 and 0.25. source
- Guardrail limitation
- Guardrail calculations require a retirement year and a positive liquid portfolio. source
- Guardrail speed
- The README estimates a guardrail computation at roughly 20–45 seconds. source
- Hosted mode
- An optional multi-user mode supports email/password accounts, server-side sessions, and profiles and scenarios stored in driftcurve.db. source
- Household modeling
- The model supports multiple people with separate retirement dates and savings rates. source
- Installation
- The quick start creates a Python virtual environment, installs requirements and runs Uvicorn on localhost:8000. source
- Intended use
- The README presents driftcurve for household retirement planning, including financial-independence planning scenarios. source
- Licence
- The project is released under the Apache License 2.0. source
- License
- The project is open source under the Apache License 2.0. source
- Limit: no guarantee
- The licence and README disclaim warranties and liability, and the maker warns that outputs can be materially wrong. source
- Limit: optimistic history
- The README characterizes the 1966–2024 period as still optimistic, with a median 90%-equity/10%-bond blended CAGR of about 10.2% annually. source
- Local privacy
- In default local mode, household information and saved scenarios remain in browser localStorage and are not written to a database. source
- Market series
- The historical simulation covers S&P 500, 10-year Treasuries, residential real estate, CPI and Bitcoin returns. source
- Method context: cash flows
- Kitces explains that Social Security, pensions, rental income, part-time work and property changes can materially alter portfolio withdrawals. source
Checked against the maker's pages on 15 Oct 2026.