Aliquot

Why I built it

A chemist friend was venting about her lab work. Not the chemistry, the arithmetic around it: unit conversions, dilutions, the same few formulas again and again. None of it is hard. It is just tedious, and there is a lot of it, and being good at chemistry does nothing to make it go faster.

So Aliquot does that part for you. You tell it what you have and what you want, and it hands back the protocol. There is no language model in the loop, on purpose. Once a problem is a formula, a model is the wrong tool: slower and less sure than the equation it would be reciting.

How I built it

The rule is boring by design. If a formula can answer it, a formula answers it, and you see every step, not a black-box number. You type your concentrations, volumes, or masses in plain language and get back a preparation protocol with the hazards flagged. It runs offline, needs no signup, and never sends your inputs anywhere. A dilution does not need the internet.

Around that I built the things you actually reach for at a bench: calculators for the cases you repeat, unit conversion, a scan step, a cookbook of common preparations, an inventory, and a warning layer that flags hazards before you pour. The work went into covering the calculations chemistry students really do, and making the input fast.

It started as a favour. She used it, it saved her the tedium, and then her friends started asking for it. That is when it turned from a script into something worth building properly, so we are doing that together. It is hers as much as mine.

Where it is

Private beta right now: onboarding students and fixing whatever they tell me is broken. Going live around October 2026.

Cheers.