Test your portfolio against
150 years of historical data

Taxes, broker fees, currency risk, withdrawal strategies
Designed to work with AI - just ask in plain English

Use it on ChatGPT

Ask about your portfolio in plain English
Answered in seconds, no code needed

Open in ChatGPT →

Free · Thinking model recommended

MCP Server

Cursor, Windsurf, AI agents

HTTP API

Python, JS, any language

Coming soon

See it in action

One question. Full analysis in seconds

everinvest.

Pre-computed examples · Real results vary by starting window

How it works

Describe your portfolio

Tell the AI your allocation, country, and goals in plain English. Or use JSON if you prefer

AI runs the backtest

150+ years of rolling windows. Real taxes, broker fees, currency conversion, and inflation all modeled

Get actionable results

CAGR, drawdowns, Sharpe ratios, plus which exact ETFs to buy with ISINs and total cost

Years of data

Unique assets

Countries with tax rules

Stress test scenarios

What's inside

Everything the backtester supports - click to explore

FAQ

Is it really free?

Yes. The Custom GPT and MCP server are free. If you wish to use the Custom GPT, you will need a ChatGPT account (free or paid - a paid account with the latest thinking model gives the best results).

Where does the data come from?

Real historical returns from peer-reviewed academic sources: Jordà-Schularick-Taylor Macrohistory Database, Fama-French Research Data Library, Robert Shiller, Dimson-Marsh-Staunton, FRED, and official government bond databases. Every backtest result includes its exact data sources - just ask.

Is this financial advice?

No. This is an educational backtesting tool. Past performance does not guarantee future results. All results are hypothetical. Consult a licensed financial advisor for personal investment decisions.

How accurate are the tax calculations?

Tax rules are modeled from official sources and updated for 2025/2026 rates. They cover capital gains, dividends, withholding taxes, special regimes (Non-Dom, NHR, Beckham Law), exit taxes, and accumulating vs. distributing ETF treatment. Always verify with a local tax advisor.

Can I use it with Cursor, Windsurf, or my own AI app?

Yes. The MCP server works with any client that supports remote MCP connections - Cursor, Windsurf, and other MCP-compatible tools. Just add the server URL to your MCP config. An HTTP API for direct integration is coming soon.

Start backtesting in 30 seconds

No extra sign-up. No code required. Just ask ChatGPT about your portfolio

Open in ChatGPT →

· HTTP API - coming soon

Connect via MCP

Create mcp.json in your config folder:

{
  "mcpServers": {
    "everinvest-portfolio-backtester": {
      "url": "https://api.everinvest.ai/mcp"
    }
  }
}

Works with Cursor, Windsurf, and any MCP-compatible client.

HTTP API

Drop this file into your project. No API key, no setup. pip install requests

everinvest.py

"""
everinvest Portfolio Backtester - Python client.
Drop this file into your project: from everinvest import backtest
"""
import requests

BASE = "https://api.everinvest.ai"

def backtest(portfolios, **kwargs):
    """
    Run a historical portfolio backtest (1871-2025).

    Args:
        portfolios: Allocation dict or list of portfolio dicts.
            Single:  {"stocks.usa": 0.6, "bonds.usa.gov.long": 0.4}
            Compare: [{"name": "A", "allocation": {...}},
                      {"name": "B", "allocation": {...}}]
            Bonds:   {"bonds.usa": {"weight": 0.4, "strategy": "ladder"}}

    Keyword args (all optional):
        investor_country:    "USA","DEU","GBR","CHE","CYP",... (default USA)
        window_years:        Rolling window length, min 5 (default 30)
        start_year/end_year: Data range (default 1871-2025)
        initial_capital:     Starting value (default 10000)
        initial_capital_currency: "USD","EUR","GBP","CHF",...
        currency_basket:     Multi-currency weights summing to 1.0
                             {"USA": 0.25, "DEU": 0.25, "CHE": 0.25, "GBR": 0.25}
        rebalance_config:    {"frequency_years": 1, "drift": {"stocks.*": 0.10}}
        withdrawal_config:   Decumulation. Strategies: fixed_pct, vpw,
                             guyton_klinger, rmd, 1_over_n, dividends_only
        tax_config:          {"enabled": True, "account_type": "taxable",
                             "other_income": 80000, "regime": "non_dom"}
        fee_config:          {"broker": "IBKR_tiered"} or custom fees
        scenarios:           ["stagflation","gfc_2008"] or custom objects
        cash_flows:          [{"amount": 5000, "recurring": True}]
        prefer_accumulating: True for acc ETFs, False for dist (default True)
        detailed:            Include all year-by-year data (default False)
        include_cagr_matrix: CAGR by start_year × period (default True)
        income_tracking:     Track dividends/gains per year (default False)

    Returns dict with: portfolios, test_parameters, country_results, aggregate

    Full docs: get_resource("guide") | get_resource("examples")
    Asset IDs: get_resource("assets") | list_assets("DEU")
    """
    r = requests.post(f"{BASE}/backtest", json={"portfolios": portfolios, **kwargs})
    r.raise_for_status()
    return r.json()

def list_assets(country="USA"):
    """Get asset IDs, ETF tickers, and expense ratios."""
    r = requests.get(f"{BASE}/assets", params={"investor_country": country})
    r.raise_for_status()
    return r.json()

def get_resource(name):
    """Fetch docs. name: guide|examples|assets|scenarios|taxation|account-types"""
    r = requests.get(f"{BASE}/resources/{name}")
    r.raise_for_status()
    return r.json()["text"]

# ── Examples ──
if __name__ == "__main__":

    # Simple 60/40
    r = backtest({"stocks.usa": 0.6, "bonds.usa.gov.long": 0.4})

    # German investor, EUR perspective, with taxes
    r = backtest(
        {"stocks.dev": 0.7, "bonds.deu.gov.long": 0.3},
        investor_country="DEU",
        tax_config={"enabled": True, "other_income": 60000},
    )

    # Multi-currency digital nomad with bond ladder
    r = backtest(
        {"stocks.dev": 0.6, "bonds.eur": {"weight": 0.4, "strategy": "ladder"}},
        currency_basket={"DEU": 0.33, "CHE": 0.33, "GBR": 0.34},
    )

everinvest.ts

/**
 * everinvest Portfolio Backtester - TypeScript client.
 * import { backtest, listAssets, getResource } from "./everinvest"
 */
const BASE = "https://api.everinvest.ai";

/**
 * Run a historical portfolio backtest (1871-2025).
 *
 * @param portfolios - {"stocks.usa": 0.6, "bonds.usa.gov.long": 0.4}
 *   Or array: [{name: "A", allocation: {...}}, ...]
 *   Bond strategies: {"bonds.usa": {weight: 0.4, strategy: "ladder"}}
 *
 * @param options - All optional:
 *   investor_country    "USA"|"DEU"|"GBR"|"CHE"|"CYP"|...
 *   window_years        Rolling window (default 30, min 5)
 *   start_year/end_year Data range (1871-2025)
 *   initial_capital     Starting value (default 10000)
 *   currency_basket     {USA: 0.25, DEU: 0.25, CHE: 0.25, GBR: 0.25}
 *   rebalance_config    {frequency_years: 1, drift: {"stocks.*": 0.10}}
 *   withdrawal_config   Strategy: fixed_pct|vpw|guyton_klinger|rmd|...
 *   tax_config          {enabled: true, other_income: 80000}
 *   fee_config          {broker: "IBKR_tiered"}
 *   scenarios           ["stagflation","gfc_2008"] or custom objects
 *   cash_flows          [{amount: 5000, recurring: true}]
 *   detailed            Include year-by-year data
 *   Full docs: getResource("guide") | getResource("examples")
 */
export async function backtest(
  portfolios: Record<string, number | object>,
  options: Record<string, unknown> = {}
) {
  const res = await fetch(`${BASE}/backtest`, {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({ portfolios, ...options }),
  });
  if (!res.ok) throw new Error(`API error: ${res.status}`);
  return res.json();
}

export async function listAssets(country = "USA") {
  const res = await fetch(`${BASE}/assets?investor_country=${country}`);
  return res.json();
}

/** Fetch docs: guide|examples|assets|scenarios|taxation|account-types */
export async function getResource(name: string) {
  const res = await fetch(`${BASE}/resources/${name}`);
  return (await res.json()).text;
}

// ── Examples ──

// Simple 60/40
await backtest({ "stocks.usa": 0.6, "bonds.usa.gov.long": 0.4 });

// German investor with taxes
await backtest(
  { "stocks.dev": 0.7, "bonds.deu.gov.long": 0.3 },
  { investor_country: "DEU", tax_config: { enabled: true, other_income: 60000 } }
);

// Multi-currency with bond ladder
await backtest(
  { "stocks.dev": 0.6, "bonds.eur": { weight: 0.4, strategy: "ladder" } },
  { currency_basket: { DEU: 0.33, CHE: 0.33, GBR: 0.34 } }
);

terminal

# Simple 60/40 backtest
curl -X POST https://api.everinvest.ai/backtest \
  -H "Content-Type: application/json" \
  -d '{"portfolios":{"stocks.usa":0.6,"bonds.usa.gov.long":0.4}}'

# List available assets for a German investor
curl https://api.everinvest.ai/assets?investor_country=DEU

# Get the quick-start guide
curl https://api.everinvest.ai/resources/guide

OpenAPI Spec ↗ Swagger UI ↗

No API key · No rate limit · OpenAPI 3.1

Test your portfolio against150 years of historical data

Use it on ChatGPT

MCP Server

HTTP API

See it in action

How it works

Describe your portfolio

AI runs the backtest

Get actionable results

What's inside

FAQ

Start backtesting in 30 seconds

Connect via MCP

HTTP API

Test your portfolio against
150 years of historical data