thai-addr

Parse Thai postal addresses into structured fields for DBD, DGA, and form ingestion workflows.

import { parseAddress } from 'thai-addr';

const parsed = parseAddress(
  '99/9 หมู่ที่ 5 ซอยสุขุมวิท 31 ถนนสุขุมวิท แขวงคลองตันเหนือ เขตวัฒนา กรุงเทพมหานคร 10110',
);

console.log(parsed);

{
  houNum: '99/9',
  bdgNme: '',
  soiNme: 'สุขุมวิท 31',
  mooNum: '5',
  steNme: 'สุขุมวิท',
  tmbNme: 'คลองตันเหนือ',
  ampNme: 'วัฒนา',
  prvNme: 'กรุงเทพมหานคร',
  posCde: '10110'
}

Install

npm install thai-addr

API

parseAddress(address, options?)

Returns a ParsedAddress object:

| Field | Meaning | | --- | --- | | houNum | House number | | bdgNme | Building, village, room, floor, or other leftover place text | | soiNme | Soi, alley, lane, or junction | | mooNum | Moo number | | steNme | Street or road | | tmbNme | Subdistrict / tambon / khwaeng | | ampNme | District / amphoe / khet | | prvNme | Province | | posCde | Postal code |

Options:

type ParseAddressOptions = {
  preservePrefix?: boolean;
  profile?: 'generic' | 'dbd' | 'dga' | 'thaid';
};

When preservePrefix is true, fields keep labels such as ถนน, ตำบล, อำเภอ, and จังหวัด.

Profiles tune how much the parser should infer from unlabeled text:

| Profile | Use case | | --- | --- | | generic | Default balanced parser for general Thai address strings | | dbd | Conservative parsing for DBD-style business address data | | dga | Conservative parsing for government structured/reference data | | thaid | More helpful inference for person/form style addresses, including unlabeled Bangkok tails such as บางนา บางนา กรุงเทพ |

parseAddress('12/34 อาคารตัวอย่าง ซ.ตัวอย่าง7 บางนา บางนา กรุงเทพ 10260', {
  profile: 'thaid',
});

Notes

Thai addresses are semi-structured. This package uses deterministic parsing rules and is best suited for data cleanup, import pipelines, and forms where the original source is already a Thai address string. For maximum accuracy at scale, validate parsed province, district, and subdistrict names against an official administrative-area dataset.