The New Zealand Gazette, "official Government newspaper and authoritative journal of constitutional record, published since 1841" (DIA) contains a remarkable variety of subjects, such as:
- registers of government employees, teachers, engineers, electricians and surveyors
- electoral rolls and lists of voting places, candidates and election results
- bankruptcy, company and partnership dissolution notices
- Native Land Court notices
- notices of land taken for roads, railway lines and schools
This project aims to make the Gazette more accessible to genealogists, historians, and everyone else who wants to learn more about the history of government decisions and communications that can be found in the NZ Gazette.
Browse Issues by Year
New Zealand Gazette
Nelson Provincial Gazette
No issues currently available
Index Statistics
178
Years Covered
24,338
Issues Processed
507,917
Pages Transcribed
8,377,981
Names Identified
2,157,807
Unique Names
1,542,583
Named Officials
101,743
Unique Officials
Questions and Answers
Why is a separate Index project necessary?
The New Zealand Gazette (along with the Provincial Gazettes) is an extraordinarily rich source of historical information and has been carefully scanned, catalogued and made available online (huge shout-out to NZLII and Victoria University Library), but limitations in the accuracy of the OCR of the scanned pages and the lack of a user-friendly search function meant that finding information within the issues was not easy; nor was gaining a 'bird's-eye view' of the content of issues in a particular year.
Ideally the Gazette would be searchable via Papers Past, but this is not currently possible as "with limited resourcing, we tend not to duplicate content already available online."
This project seeks to act primarily as a 'finding aid' for both historical and recent issues of the NZ Gazette, promoting access to the information in the Gazette in both detail (finding individual names) and broad strokes (eg, summaries of Supplement and Extraordinary issues).
The initial project motivation was producing an accurate nominal index of persons named in the NZ Gazette. For genealogical research, a nominal index is significantly more useful than using a generic text search to find individual names, as they may be recorded in numerous variations. The fresh transcription of the Gazette issues allowed for a more accessible, easy-to-use full-text search functionality as well.
As the project progressed it became clear that a high-level index of single-purpose issues would also be useful, as well as a way to easily find specific types of content within the corpus (eg, lists of vaccinators, teachers, nurses, marriage celebrants, and so on).
How much of this site is AI Generated Content?
To index the enormous amount of text within the NZ Gazette would be a truly daunting task without the help of modern LLM ("AI") tools. All searchable data, Markdown transcriptions, and issue/page summaries were generated using β¨ AI (large language models). While I strive for accuracy, you should verify critical information against the original source PDFs linked on each search result / page.
The majority of page transcription to Markdown was done using the Qwen3-VL-235B-A22B-Instruct model, and a smaller number with Qwen2.5-VL-72B-Instruct. Pages with more complex layouts such as large tables were transcribed with Gemini-3-Flash and Gemini-3-Pro.
Content analysis (which includes nominal indexing) was done with a variety of models, primarily MistralAI's Devstral-2-123b-Instruct-2512 which proved to be very reliable. Google Gemini-2.5-Flash and Gemini-2.5-Flash-Lite were also used for some pages.
Models were accessed via several inference providers: Chutes, Nvidia, ElectronHub, OpenRouter, and the Gemini API.
How were the names indexed?
The nominal index was created by having an AI model read the transcribed Markdown text of an individual page (with some context of the previous page included) and analyse the text, identifying distinct 'notices' (eg, a bankruptcy notice, or a proclamation) and identifying the individual names within each notice. For each name, the model recorded:
- surname
- forenames (including initials)
- any titles (eg Esquire, Dr)
- the 'context' in which the name is mentioned (eg, a creditor in a bankruptcy notice, or being awarded a military promotion)
Government Officials
The indexing process makes a distinction between the named individuals who were the direct subjects of a notice and the government officials who approved or posted the notice. The officials are recorded in a separate index from the main nominal index.
How was the page text transcribed?
One of the main challenges of this project was the need for a fresh transcription of the text of each page (OCR) in the source documents (pre-2000). This was necessary for accuracy as the PDF OCR had wildly varying levels of accuracy both on a character/word level, and in terms of layout recognition. These issues made it impossible to perform accurate searches of the Gazette text using the PDF sources.
Doing a fresh transcription also avoids potential copyright issues where third parties may hold copyright over the OCR text in the source PDFs (which may be considered a separate work from the raw images of the Gazette pages).
The transcription was done using AI models to generate a version of the page in Markdown format. This allows for much of the original formatting in the Gazette (eg, headings, italics, and tables) to be represented in a text-based format. Rather than traditional OCR which simply identifies all of the words on the page, knowing which parts of the text are headings and which words are in particular cells of a table makes AI content analysis more accurate.
The transcription output is by no means perfect, and different pages are not necessarily consistent in their formatting. Many tables are not formatted correctly, for example. But overall I am satisfied it is a substantial improvement over the previously available PDF OCR text.
Note: Gazette issues after ~2000 are born-digital documents, so their PDFs did not need to undergo an OCR step. This simplifies the transcription process immensely as the model just has to add Markdown formatting.
What are these little icons everywhere?
Page content, notices and single-purpose issues have been categorised (by AI) into 'domains' or broad areas of government, represented by icons throughout the website:
- ποΈ Governance & Central Administration
- π° Finance & Revenue
- βοΈ Justice & Law Enforcement
- π External Affairs & Territories
- πͺΆ MΔori Affairs
- πΊοΈ Lands, Settlement & Survey
- πΎ Primary Industries & Resources
- ποΈ Infrastructure & Public Works
- π Transport & Communications
- π Trade, Customs & Industry
- π‘οΈ Defence & Military
- ποΈ War-Related Services & Veterans' Affairs
- π₯ Health & Social Welfare
- π· Labour & Employment
- π Education, Culture & Science
- ποΈ Housing & State Construction
- π’ State Enterprises & Insurance
- π Immigration
- ποΈ Provincial & Local Government
- π¨ Emergency Management
- π° NZ Gazette
These domains are obviously subject to interpretation and have some overlap, but are useful to help quickly distinguish between, for instance, a list of teacher registrations and a list of WWII draft ballots.
What about living people?
This project includes recent issues of the NZ Gazette (up to 2017) which contain the names of many living people in a variety of contexts. Inclusion of a name in a list reflects only the information published in that specific issue of the Gazette at that time, and does not necessarily reflect the current status of that individual.
You should always refer to the linked source PDFs to verify information, and consult subsequent issues of the Gazette for possible corrections or errata.
Most prominent officials (by Gazette notice)
| Official | Notices | Position |
|---|---|---|
| William Stanley Goosman (1890-1969) | 18,093 notices | Minister of Works, Transport, Marine, Housing, Railways in the First National Government from 1949 to 1954 (Wikipedia) |
| Frank David Thomson (1877-1934) | 17,176 notices | Clerk of the Executive Council and Secretary of the Prime Minister's Department from 1926 to 1934 (Archway) |
| Percy Benjamin Allen (1913-1992) | 16,200 notices | Minister of Works in the Second National Government from 1963 to 1972 (Wikipedia) |
| Thomas James Sherrard (1895-1988) | 14,812 notices | Clerk of the Executive Council and Secretary to the Cabinet from 1948 to 1960 (photo) |
| Sir Charles Fergusson, 7th Baronet (1865-1951) | 10,818 notices | 3rd Governor-General of New Zealand from 1924 to 1930 (Wikipedia) |
| Cecil Albert Jeffery (1888-1970) | 10,768 notices | Clerk of the Executive Council and Secretary to the Cabinet from 1935 to 1946 (Papers Past) |
| William Lambert Young (1913-2009) | 10,453 notices | Minister of Works and Development in the Third National Government from 1975 to 1981 (Wikipedia) |
| James Frank Andrews (1848-1922) | 9,835 notices | Clerk of the Executive Council from 1909 to 1919 (Cyclopedia) |
| Arthur William de Brito Savile Foljambe, 2nd Earl of Liverpool (1870-1941) | 8,520 notices | 16th Governor and 1st Governor-General of New Zealand from 1912 to 1920 (Wikipedia) |
| Bernard Cyril Freyberg, 1st Baron Freyberg (1889-1963) | 8,397 notices | 7th Governor-General of New Zealand from 1946 to 1952 (Wikipedia) |
