Skip to content
CoffeeHQ

Tools

Every one of these shows its working, says what it assumes, and ends somewhere rather than stopping at a number.

Make a recipe

How much coffee, how much water, how much milk?

Quantities before brewing, in whichever direction you have the numbers. Every one of these shows its arithmetic so you can do it yourself next time.

Improve my coffee

Something is wrong, or could be better — what do I change?

These take a symptom rather than a quantity and rank what to change, one thing at a time. They give ranked guidance rather than a diagnosis, because coffee faults rarely have exactly one cause.

Plan and buy

What should I buy, and how much coffee should I keep?

Category-level decisions and household planning. No brands, no prices, and several of these will tell you that you do not need the thing you were about to buy.

Understand

What does this actually mean?

Reference and interpretation rather than a recipe — including the tools whose honest answer is a range rather than a number.

Start from a question, not a tool

Each route takes several tools and the pages between them in order. Every tool also says, under its result, what to do next.

I have coffee and a brewer

You own something to brew with and want a good cup from it.

Find your brewer's page, start from its recipe, and change one thing at a time. The later steps are for when the cup is not right, and for finding out why a change worked.

  1. Find your brewerWhat it does to the coffee, and a starting recipe labelled as a starting point.
  2. Work out the amountsCoffee and water for the cup you are making; the recipe scaler follows from it.
  3. If the cup is not rightChoose the brewer, tick what you noticed: the quickest check, and one change.
  4. One change per brewThe same loop written out for filter coffee, from the first brew to a recipe you keep.
  5. Why the change worksStrength and extraction as two numbers, and what a taste can and cannot tell you.
  6. Test it yourselfHow to compare two cups so the difference is the thing you changed.
  7. See the variables togetherWhat each control moves, side by side, before you touch the next one.

My espresso tastes wrong

You have a machine and a grinder, and the shot is sour, bitter, thin or unpredictable.

Start from what you noticed and not from a rule. The order here is the order to rule things out in: the recipe, the puck, the grinder, the water, the coffee.

  1. Read the shotEvery espresso reading: what to check first, and the one thing to change.
  2. Dial in, one change per shotThe loop from a first shot to a recipe, with what to hold still each time.
  3. The puckDose, distribution and tamp in order, and how to tell which one moved.
  4. The grinderWhen it will not go fine enough, one step is too large, or yesterday's setting is off.
  5. The waterWhen every coffee tastes flat or sharp whatever you do to the recipe.
  6. The coffee itselfVery fresh, stale, or a roast that brews unlike the last one.
  7. Enter the shotDose, yield, time and taste in; ranked suggestions out, with what to watch next.
  8. Why the puck behaves as it doesWhat was measured about pressure, flow and channelling, and by whom.

I do not know what brewer to use

You have not chosen a brewer, or wonder whether a different one would suit you better.

Start from what you want in the cup and what you will put up with, not from a ranking. No method is ranked above another here; each one trades something for something.

  1. Start from what you wantEach want, the methods that lean towards it, and what it costs you.
  2. Set two or three side by sideBody, clarity, strength and the rest, with what each statement rests on.
  3. Read the method's own pageWhat the brewer does, and its starting recipe.
  4. What it is like to ownSize, cleaning, what else you need, and faults that come from the design.
  5. Work out the first brewAmounts for the starting recipe, at the size you are making.
  6. Then make it betterOnce you are brewing: by brewer, by what you noticed.

How these work

Everything runs in your browser. Nothing you type is sent anywhere, nothing is stored, and no tool needs an account. Where a result depends on evidence with real spread it is shown as a range and labelled as one, and where it depends on judgement rather than arithmetic the page says so rather than presenting a heuristic in the same voice as a multiplication.