What does "validated" mean on this page?
passed — the dataset passed the integrity, referential-consistency and per-family coverage checks the build runs before any page is allowed to publish.
Checked against Palworld v1.0.3 · verified 2026-08-25 · Data status · Sources · Methodology
The live dataset behind this site is Palworld v1.0.3, validation status passed, with claims last verified 2026-08-25. It covers 299 Pals and currently supports 1411 published pages, each traceable to a graded source rather than to an unverified guess.
| Version | Pals | Parent pairs | Same-species-only | Passives | Active skills with a route | Obtainable Skill Fruits | Pal locations known |
|---|---|---|---|---|---|---|---|
| v0.7.0 | 227 | 25,878 | 15 | 92 | 235 of 235 | 89 | 0 of 227 |
| v1.0.3 (current) | 299 | 44,850 | 26 | 115 | 312 of 320 | 91 | 147 of 299 |
"Active skills with a route" counts the skills the engine can say something about — the ones reached by a learnset entry, an obtainable Skill Fruit or a verified acquisition rule. The second number is every active skill in the dataset. The difference is skills that exist with no verified way to get them on any Pal; they are listed on the all-active-skills page and answer "needs verification" rather than "no".
"Pal locations known" is a hard limit of the source, not of this site: the wiki's location table covers a named-area spawn for some Pals and nothing for the rest. A Pal profile shows a "Where to find" section only when a location is known — the other Pals show no such section, never a placeholder or a guess.
A family marked partial means we know our coverage has gaps, and the engine answers "needs verification" instead of "no" for anything that depends on it.
| Family | v0.7.0 | v1.0.3 |
|---|---|---|
| breeding | complete | complete |
| passives | complete | complete |
| active skills | complete | complete |
| active skill exclusivity | missing | missing |
| learnsets | complete | complete |
| items | complete | complete |
| mechanics | complete | complete |
The same files the pages on this site are built from, keyed 1.0.3@2026-08-12-29cd766c — game version and dataset revision — so a copy you pull
today keeps answering the same way after this site promotes a newer one. Every record
carries the claim id it rests on, and those ids resolve on the evidence page, so a
disagreement can be traced to a source rather than argued about.
| File | Licence | Attribution required |
|---|---|---|
| Breeding edges | MIT | PalCalc (tylercamp and contributors), MIT — https://github.com/tylercamp/palcalc |
| Passive skills | CC BY-SA 4.0 AND MIT | PalCalc (tylercamp and contributors), MIT — https://github.com/tylercamp/palcalc; palworld.wiki.gg contributors, CC BY-SA 4.0 — https://palworld.wiki.gg |
| Learnsets | CC BY-SA 4.0 | palworld.wiki.gg contributors, CC BY-SA 4.0 — https://palworld.wiki.gg |
| Pal index | MIT | PalCalc (tylercamp and contributors), MIT — https://github.com/tylercamp/palcalc |
Licences differ by file because the sources do: breeding edges and fixed passives come from PalCalc, learnsets from the Palworld wiki. Re-using the CC BY-SA 4.0 parts means attributing them and sharing alike. The full notice ships beside the files as LICENCE.txt.
Items, locations and the tier lists are deliberately not in the dump. They rest partly on sources that publish their data openly but license nothing, and a dump is redistribution in a way a page citing a fact is not. The build enforces this rather than trusting it: each file is walked for the claims it cites, and a claim reaching an unlicensed source fails the build instead of shipping. If you need those, the pages that show them are the answer, and the JSON endpoints serve the same facts one question at a time.
Species exclusivity for active skills is not populated: no source we have pinned exposes the flag. The engine supports the rule, and no active-skill page has been frozen while it is open. Where exclusivity would have been the reason, the answer still comes out correct through the other channels — with a less precise explanation.
Historical coverage is one version deep (v0.7.0). A build that was only obtainable before that has no dataset to prove it, so it answers "needs verification" rather than "no".
passed — the dataset passed the integrity, referential-consistency and per-family coverage checks the build runs before any page is allowed to publish.
2026-08-25, against the v1.0.3 dataset generated 2026-08-12. That date is the same freshness stamp shown on every published page.
The engine answers "needs verification" instead of "no" for anything that depends on it, rather than guessing — see the completeness table below for exactly which families are affected.