Changelog
What changed on this site, when, and what justified it. A density is never edited without a published reference and an entry here, and corrections sent in by readers are recorded with credit. If a figure on this site looks wrong to you, tell us and it ends up on this page.
Every figure now says where it came from, including where it did not
The section at the bottom of each ingredient page used to be the same paragraph 48 times, naming USDA FoodData Central and the King Arthur Baking ingredient weight chart under every ingredient regardless of whether either had been checked for it. That read like a citation without being one, and in the case of all-purpose flour it named King Arthur as a reference for 125 grams per cup when King Arthur publishes 120. Each page now shows a table of the references actually consulted for that specific ingredient, the figure each one publishes, our figure alongside them, and a plain explanation of the difference. Ingredients we have not yet sourced say so in as many words.
All-purpose flour: the disagreement with King Arthur is now stated openly
No figure changed. The site continues to use 125 grams per US cup. What is new is that the page now states that King Arthur Baking publishes 120 grams for the same measure, explains that the gap comes from their stricter spoon-and-level method, and says which figure to use for which recipe.
Rolled oats checked against King Arthur
King Arthur Baking publishes 89 grams for a cup of old-fashioned rolled oats. This site prints 90. The two are the same figure for kitchen purposes and neither was changed, but the page now shows both so the one gram is not a mystery.
Ounce columns relabelled as weight, and a page explaining why
The ounces column in every conversion table sat directly beside a millilitres column, which made it read as fluid ounces. It was never fluid ounces. Every column is now labelled as ounces of weight, every table carries a line explaining the difference, and there is a new page working through the whole confusion with the weight of one cup of each ingredient against its fixed 8 fluid ounces.
Two pages disagreed with the rest of the site about the same number
The cups-to-grams guide rounded half a cup of flour to 63 grams while the flour page printed 62.5, and the cups-to-ml guide rounded a third of a cup to 5 tablespoons and two thirds to 11, where the ingredient pages correctly printed 5.33 and 10.7. Both pages were rounding independently instead of using the site-wide formatter. They now use it, the rule is written down on the methodology page, and there are tests that fail if either drifts again.
The ingredient count in llms.txt was wrong
The machine-readable summary at /llms.txt claimed 45 ingredients where the site has 48, and pointed at the sitemap for the full list. It is now generated from the same data as the pages, so the count cannot drift again, and it points at the new ingredient weight chart instead.
Three ingredients added on Search Console demand
Sweetened condensed milk, vegetable shortening and chopped dates were added after Search Console showed real queries for each landing on the homepage with no page to answer them. Each figure was checked against a named reference before the page was created.