A live, editable portfolio workbench. It opens with a sample book priced on real quotes through June 3, 2026, replace it with your own holdings, in any currency, then refresh prices, model a rebalance, run what-ifs, record sales, and ask the AI of your choice about anything you are unsure of.
🔒 Private by design — your data never leaves your browserEditable. Change shares, cost, price, currency, purchase date, or yield and everything below recomputes. Buy a ticker more than once and the lots roll up into one row, tap the arrow to expand, edit, or add lots. Purchase dates feed the holding-period and tax math; the sample uses real January 2024 entry and current prices.
The AI features: fetching live prices, running custom what-ifs, and answering your questions in the "Ask the AI" box. Pick an engine on the right.
• Claude works right here with no setup. Just use the features.
• Which model? Haiku is the recommended default for everyday use, fetching prices, adding benchmarks, and answering general questions, and it costs the least. Switch to Sonnet when you want deeper reasoning: tax-lot optimization, rebalancing decisions, or thinking through concentration risk. The same tiers exist for OpenAI (4o-mini vs 4o) and Gemini (Flash vs Pro).
• To use OpenAI or Gemini instead, select it, then paste an API key. A key is a free private code you create on the provider's website:
OpenAI platform.openai.com Gemini aistudio.google.com Anthropic console.anthropic.com
Your key stays on your device, is sent only to the engine you pick, and is never saved. Claude and Gemini can look up live web data; OpenAI answers from what it already knows.
| Ticker i | Name i | Class i | Shares i | Cost/sh i | Price i | Ccy i | Bought i | Yield i | Value i | Return i | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Total Book | |||||||||||
The investor's worry, quantified. This book is not really diversified: it is a single-country, mostly-mega-cap-tech bet wearing fourteen tickers.
Concentration is invisible while it is working. The cost shows up as the rallies you were not in. 2025 was exactly that year.
Add a hypothetical investment and it draws as a new dollar line on the chart above. Tap a preset, or describe your own and the selected model fetches the history. Then use "add to book" to blend it into your totals instead.
The investor asked specifically about commodities and international stocks. Here is the case for each, and the case against chasing them.
Gold has been the best-performing major asset class of the past two years. It started 2025 near $2,624 an ounce and peaked around $5,589 in late January 2026 before easing back toward $4,500. The drivers are structural: heavy central-bank buying from China, India, and Turkey, a softer dollar, and demand for a hard asset amid geopolitical stress including the Iran conflict.
Why it diversifies: commodities, and gold in particular, tend to have low or even negative correlation to stocks. They behave like insurance against inflation and geopolitical shocks, the scenarios that hurt a stock-and-bond book most. This portfolio holds exactly zero.
The caution: gold pays no dividend or interest, it can be volatile, and after a near-double it is not cheap. The case for adding it rests on its low correlation going forward, not on its trailing return. A typical diversifying weight is roughly 5% to 10%, via a broad gold or commodity fund.
After a decade of US dominance, foreign equities led in 2025. Developed markets (MSCI EAFE) returned about 32% and emerging markets about 32%, both roughly double the S&P. The catalysts were a rotation from growth into value, a weaker dollar, far cheaper valuations abroad, European fiscal and defense spending, and the AI boom spreading into the Asian chip supply chain.
Why it diversifies: this book is 100% US, so every position depends on one economy, one currency, and one policy regime. Adding developed and emerging markets spreads that across dozens of countries and currencies, at valuations well below US levels.
The caution: international carries currency risk, weaker governance in some markets, and concentrated country exposure in EM, notably China. And again, you would be adding after a strong run. The durable argument is the valuation gap and the diversification, not the recent scoreboard.
Drag the sliders to fund a commodities and international sleeve out of the concentrated US equity book. Fixed income is held steady. Everything recalculates from your live holdings.
What the book pays you, and how its gains would be taxed if sold today, shown in your display currency. Driven by the yield, currency, and purchase-date fields in your holdings.
Unrealized gains and forward income are estimates based on today's prices, yields, and FX. Yield on cost compares today's income to what you originally paid. Not tax advice.
A 12-month projection of the dividends and interest your current holdings would pay, distributed by typical schedule (equities quarterly, bonds and cash monthly, commodities annually), in your display currency.
Sell part or all of a ticker. Choose how lots are drawn down, FIFO, LIFO, average cost, or a specific lot. Proceeds convert at the exchange rate on the sale date, cost at each lot's purchase-date rate, so realized currency gain is captured too. Selling reduces shares; removing a sale puts a lot back.
Nominal return is the number everyone quotes. Real return is what your money actually bought you after inflation. This deflates each lot by US CPI from its purchase month to today, the way institutions report performance and most retail tools never do.
Real return = nominal return with inflation removed. Your key is stored only on this device. CPI is the US CPI-U series (CPIAUCSL); real returns reflect US prices and are an approximation for non-US holdings.
Runs on your chosen model, set by the selector above the holdings table: Claude (works in-app with no key), or OpenAI / Gemini with your own key. Claude and Gemini fetch live web data; OpenAI answers from its training. It sees your current holdings, purchase dates, yields, and your real (inflation-adjusted) return, and accounts for stock splits.
This tool has no server, no account, and no sign-in. Everything you type or import, your holdings, cost bases, purchase dates, and yields, lives only in your own browser. When self-hosted at folio.icehunter.net it is saved to your browser's local storage on your device, so it is still there next time you visit, and it is never uploaded anywhere.
The site owner cannot see any of your numbers. There is no database receiving them. The only time anything leaves your browser is if you choose to use an AI feature, and then your request goes directly from your browser to the AI provider you selected (Anthropic, OpenAI, or Google) using your own key. It does not pass through the site owner.
If you want a clean slate, clearing your browser data for the site removes everything. Nothing is retained elsewhere.