Paste a recipe or name a dish. Every ingredient gets mapped to the organism it came from, the eukaryote tree gets pruned to what's on the plate, and the branch lengths get added up: millions of years of evolution per meal.
Dashed chips were inferred, from a dish name or from an "or" alternative in the recipe. Remove anything wrong, add anything missing. Click a name, here or on the tree, for Wikipedia, OneZoom and GBIF.
Each organism's private share of the total: branches it alone reaches. What's left over is shared trunk.
Meals you log stay in this browser. The weekly score is the PD of everything eaten in the last seven days combined, which is where the shallow branches start to matter.
Tree version open-0.18.2 (2026-09-30). The built-in tree has 0 edible taxa and 0 Myr of branch length. This build is assembled entirely from open, redistributable sources: species-level divergence times from Smith & Brown 2018 (seed plants), Nitta et al. 2022 (ferns), Rabosky et al. 2018 (ray-finned fishes), Upham et al. 2019 (mammals), Jetz et al. 2012 (birds), Stein et al. 2018 (sharks and rays), Jetz & Pyron 2018 (amphibians), Tonini et al. 2016 (squamates), Varga et al. 2019 (mushrooms), Shen et al. 2020 (ascomycete fungi) and Kim et al. 2026 (red algae; tree file provided by Hocheol Kim), hung on a backbone of deep-node ages from Strassert et al. 2021 and vertebrate nodes from Irisarri et al. 2017. Molluscs, crustaceans, insects, brown and green seaweeds, turtles, crocodilians and the small phyla have no species-level open chronogram yet; their families sit on a class- and order-level skeleton whose ages are cited where a dated phylogenomic study exists (decapods: Wolfe et al. 2019; bivalves: Li et al. 2025; cephalopods and gastropod subclasses: Tanner et al. 2017; insect orders: Misof et al. 2014, Peters et al. 2017, Kawahara et al. 2019; brown algae: Choi et al. 2024) and otherwise approximate and flagged for review (see node_review_open.csv). Placements on that skeleton join at the order's crown, which slightly overstates PD for those groups.
Dish names are resolved with a curated lookup table (version dishes-0.2.1 (2026-09-30), 0 dishes with their usual ingredients), kept as data/dishes.csv in the repository. Anything not in the table falls through to ingredient-name matching against the taxon table's aliases. Spanish works too, with regional names (choclo, elote, jojoto; palta, aguacate; poroto, frijol, habichuela) and with or without accents. Everything runs in the page; nothing is sent anywhere.
Reading the numbers. The tree is eukaryotes only, so bacteria are not scored: the cultures in natto, yogurt, kimchi, sauerkraut and cheese count for the plant or animal they are grown on, not for themselves. The root, where plants and fungi-plus-animals part, sits at 2.13 billion years in the source used here (Strassert et al. 2021); other published estimates run from about 1.5 to 2.1, so any meal with both a plant and an animal reaches back that far, and the deepest branches carry the widest uncertainty. Within groups placed on the skeleton rather than a species-level source, families join at their order's crown, which overstates shallow splits a little.
To test another tree, load a Newick whose tip labels are ids plus a CSV with columns id, common_name, species, lineage, clade, aliases; clade is one of animal, fungus, plant, green, red, brown, other; branch lengths in Myr, ultrametric.
PD is Faith's (1992) phylogenetic diversity: the sum of branch lengths of the minimum subtree connecting the meal's taxa to the root of the tree. Richness is the count of distinct taxa. Coverage divides the meal's PD by the total branch length of the tracked edible tree, so it depends on how many edible taxa are tracked. The span below the common ancestor is the same sum without the path from the meal's MRCA to the root; it is zero for a single-organism meal. Bacteria are not scored, on purpose: the tree is eukaryotes only, so the cultures in natto, yogurt, kimchi, sauerkraut and cheese count for the plant or animal they are grown on, not for themselves (a bacterial branch would be several billion years long and would swamp every meal it touched). Amounts are ignored; an abundance-weighted variant (Chao, Chiu and Jost's phylogenetic Hill numbers) is the natural next step.