Our Predictive API has been due an update for some time. We’ve been hard at it adding a bunch of new stuff to make it more configurable and useful than ever before. Let’s jump into it!

More addresses

No matter what tool, app or website you’re building, simple customer registration and address validation is front-and-centre of any good user experience. With more addresses than ever before, the API makes it easier for users to find the address they’re looking for, in fewer keystrokes.

  • Predictive suggestions of mailing addresses

Are your users looking for more than just physical addresses? No problem. The beta API now has 14.5 million mailing addresses, including:

  • PO Boxes - Post Office Boxes
  • RMB – Road-side Mail Box
  • GPO Box – Multiple PO Boxes across the country with the same number
  • Private Bag – Used in rural and remote areas
  • Locked Bag – Used by businesses receiving a lot of mail

  • Geocoded National Address File (G-NAF)

You can now select Australia’s comprehensive and trusted address data source, G-NAF, as a data source in the API, as well as our daily-updated G-NAF Live.

  • Dataset parameter

To enable these new address sources, we’ve implemented a new parameter called ‘dataset’. It’s optional. If you don’t use it, you’ll experience the same Predictive API as version 1.2.0.

If you do use it, you can make queries against G-NAF address or mailAddress. We also support the combining of G-NAF and mailAddress by way of a comma-separated list. Unfortunately, you can’t combine G-NAF Live in the same call now, but we’re working on it.

Better linkages, every building

We’ve been listening.

  • Predictive suggestions that link to our Buildings API and Addresses API

Using the new ‘dataset’ parameter and selecting G-NAF as the data source, you get predictions that feed directly into the Buildings API. It reduces the number of calls you need to make to get buildings data and simplifies your development.

Fuzzy logic on common misspellings, aliases and more

Everyone knows what an address is, but we often write them differently – whether it’s spelling mistakes, formatting or structure, or even forgetting the exact name of a suburb. Fuzzy logic helps match the addresses your users are looking for, even when they’re not 100% sure what they’re looking for.

  • Address abbreviations – Some users like to be formal and look for ‘Unit’, others just type ‘U’. We’ll match them no problem.
  • Misspellings of plural and non-plural - Handy when users can’t remember if it was Smith St, or Smiths St...
  • Locality/suburb aliases – Useful when users know an area as a suburb that doesn’t exist, for example, Manuka in the ACT or Kings Cross in NSW.
  • Number formatting – Allows for more flexibility in the formatting of a ranged number address or a unit number address.

What's deprecated

  • XML formatted responses

We'll drop support for XML formatted responses when version 1.3.0 of the Predictive API moves out of Beta.

If you’re using the XML format, have questions about the API or want to make a feature request, please contact us at beta@psma.com.au or use our support portal.


Useful Links

Developer Portal
Beta API document
Support Portal
GitHub repo