CalPal

CalPal sits in a small category that is easy to undervalue until you use it: the literate calculator. The format — prose interleaved with live math — is not cosmetic. It changes what you do with the tool. In a spreadsheet, the formula is hidden behind the cell and the reasoning lives in your head. In CalPal, the sentence "flights for four come to 1,840, plus 18% for the resort fee" IS the calculation, and the answer sits on the line where you wrote it. The engine is broader than the marketing copy suggests. Variables persist and can be reused across a whole document. Unit conversions work inline (km to miles, USD to EUR) and mixed units resolve automatically — 5 km + 2000 m comes back as 7 km. Date math handles both offsets (today + 3 weeks) and recurring calendar positions (the 4th Saturday of Feb 2026), which is the kind of question a spreadsheet makes you fight with a lookup table. Timezone math is a one-liner: 9am in Tokyo resolves against your local clock. Statistical helpers like total and mean mean you can drop a column of numbers into prose and get a figure back without leaving the paragraph. The AI assistant is where CalPal separates from its peers. Give it a prompt — "draft a travel budget for four people, five days in Lisbon, flights and hotels and daily spend" — and it produces a complete calculation set with labeled variables. That is the blank-page problem solved, and it is the feature that will keep people coming back after the novelty of plain-English math wears off. Using Google Gemini with an optional bring-your-own-key setup means you can run it on your own quota, which matters if you are pasting client figures into a document. Privacy is the quiet strength. There is no account, no server round-trip for your numbers, and storage is local to your browser. Autosave keeps your work between sessions and URL sharing hands a finished calculation to someone else without exposing an account or a workspace. For a consultant sending a client a link to a fully narrated cost breakdown, that is a real workflow that a spreadsheet cannot match. The limitations are honest and worth naming. There are no charts or graphs — CalPal is a text tool and it stays a text tool. There is no real-time collaboration; two people cannot type into the same calculation at once. There is a web interface and no indication of mobile or desktop apps, so offline work is not on the table. The AI assistant may need a user-supplied API key, which means provider rate limits apply to whatever key you bring. And the tool is a scratchpad, not a model: a shared five-year financial model with 40 tabs belongs in a spreadsheet, not here. Where CalPal fits best: solo analysts writing projections where the assumptions matter as much as the output, project managers turning a cost estimate into something a stakeholder can actually read, engineers doing conversions and cross-timezone scheduling in a single document, and students who need to show steps. Where it does not: