HS classification without the PDF tariff schedules.

This is the international Harmonized System (HS 2022) maintained by the World Customs Organization — the 6-digit base that every country shares. It is not the US HTS, the EU CN, or any other national schedule. Those national systems take the WCO 6-digit code and extend it to 8 or 10 digits (e.g. the US HTS adds digits 7–10 for duty rates). We return the universal 6-digit classification and its hierarchy; you append the national extension your customs authority requires.

Sections and chapters are complete: all 21 WCO sections (I–XXI) and all 99 chapters (01–99, with 77 reserved), titled verbatim from the HS 2022 edition. Below that we carry 131 deeper entries — 116 4-digit headings and 15 6-digit subheadings — a representative sample spread across many chapters (coffee, apparel, electronics, vehicles, machinery and more), each with correct parent and section linkage. The dataset is structured so deeper headings extend cleanly chapter by chapter.

Yes. The lookup endpoint normalizes input before matching — dots and whitespace are stripped, so 09.01.11, "0901 11" and 090111 all resolve to the same subheading. You can hand it codes straight from a tariff document without reformatting. Lookup is exact: a normalized code that is not in the dataset returns a not-found rather than a fuzzy guess.

A section is the top-level grouping, identified by a roman numeral (I–XXI) — for example Section II, "Vegetable products". A chapter is a 2-digit code (01–99) that sits inside a section — chapter 09, "Coffee, tea, maté and spices", belongs to Section II. Headings (4-digit) live inside chapters and subheadings (6-digit) inside headings. GET /sections lists each section with its chapter membership; GET /chapters lists every chapter with its parent section.

GET /search matches your query case-insensitively against entry titles and ranks results best-first: an exact whole-word match outranks a word-prefix match, which outranks a loose substring; within a tier, shorter (more specific) titles sort ahead. It searches descriptions, not code numbers — use /lookup when you already have a code. Pass an optional limit (1–100, default 25) to cap the result set.

The official HS nomenclature ships as PDF schedules that are painful to parse and easy to get wrong on parent linkage and section assignment. This API gives you the same taxonomy as predictable JSON with a stable code → entry contract — ideal for customs, landed-cost, duty-estimation and shipping tools that need to classify a product or resolve its chapter and section programmatically.