HtAG opens Australian sold-property data to AI assistants

By Nathan Brooks

Only 33 of 274 commercial sales in HtAG Analytics’ sample carried a recorded price. That left 88.0% without one. The sample covered sales within 5 km of Sydney’s CBD from 1 November 2025 through 4 November 2026.

What changed on 4 November 2026?

On 4 November 2026, HtAG launched commercial sold-property search across Australia. Its Developer Portal offers a self-serve REST API and Model Context Protocol, or MCP. At A$0.075 per row including GST, with 25 rows free on every endpoint, the service changes how teams retrieve evidence but leaves undisclosed commercial prices unresolved.

Most commercial sales use “contact agent” marketing, and prices are rarely published, according to HtAG. Victoria’s rule requiring agents to publish residential sold prices within seven days has been in force since 1 October 2026. It does not require commercial agents to publish prices.

Who is this workflow for?

At the comparable-sales stage, commercial agents, buyers’ agents, valuers, lenders, brokers, developers assessing sites and proptech teams can query sold records through one API call or assistant question. The release gives no measured cycle-time gain or results by team size. It therefore supports a workflow change, not a productivity claim.

Founder Mat Djolic said: “What changes today is the search. Comparable evidence that used to take phone calls, PDFs and spreadsheets is now one API call, or one question to an AI assistant.” He also said: “Most commercial sale prices never become public, and we would rather show that plainly than pretend otherwise.”

The market around the tool is measurable. Ray White counted A$85.6 billion in Australian commercial property sales across 9,015 transactions in 2025. CBRE reported A$19.0 billion of commercial investment in the first half of 2026, 16% more than a year earlier.

How does the sold-property API work?

The GET /v1/property/commercial/sold/search endpoint can search around an address, an address key, or latitude and longitude. It supports radii up to 5 km or a whole suburb.

Results can be filtered by:

• Nine property categories: industrial, office, retail, showroom, medical, development land, hospitality and leisure, rural agribusiness, and other.
• Tenanted investment or vacant possession.
• Sale date and sale price.
• Floor area and land area.

Users can sort by newest sale or nearest distance, and each result may include address, distance, sold date, price where disclosed, floor area, land area, zoning, land use, year built and tenure. HtAG says a recorded price represents the transfer of title. It never represents an asking price.

What did the field test show?

In a 4 November 2026 test for industrial sales within 5 km of 6 Woods Close, Huntingwood NSW, sold since 1 November 2025, the API returned 14 sales. Eight had vacant possession, four were tenanted, and two did not state tenure. Floor areas ranged from 211 to 5,248 square metres. Only one sale had a recorded price.

The test shows one returned result set. It does not establish market-wide coverage or data completeness. The release provides no independent verification for its 274-sale disclosure sample. The stated limits are geography and period.

Can an AI assistant run the same search?

HtAG also offers the search as the search_commercial_sold_properties tool on its MCP server. Users can connect HtAG to Claude or another MCP-compatible assistant. They can ask: “industrial sales within 5 km of [address] in the last 12 months, tenanted only, nearest first.” HtAG’s connectors appear in Anthropic’s Claude connector directory.

MCP is the open standard that lets assistants such as Claude call external data tools, which puts this request on the same HtAG search rather than a separate dataset. The practical effect is a plain-language query over the sold-property endpoint. It is not a new commercial transaction process. For proptech teams, the API supports embedded comparable evidence. For assistant users, the MCP tool supplies the same query layer.

What does access cost?

Access is available at htagai.com, including a no-code Playground for browser searches. Enhanced-tier list pricing charges A$0.075 per row including GST. Every account receives 25 free rows on each endpoint, while linked annual HtAG Professional and Elite plans receive 90% off data rates, and usage is capped at 10,000 billable rows per organisation each month.

HtAG says it is among the first Australian providers to expose commercial comparables to AI assistants through MCP. It also claims to be one of the few offering self-serve, pay-per-row access. It says comparable data is otherwise sold mainly through enterprise contracts. No named competitor comparison appears. The market-position claim remains unverified.

What does the product leave out?

The endpoint covers sold comparable-sales data only. It excludes commercial leasing, rents, yields and for-sale listings. It also omits sales whose address the selling agent withheld. That is a narrower record than the full commercial property market. The 12.0% recorded-price sample explains why.

HtAG serves property investors, buyers’ agents and mortgage professionals. It also serves proptech developers and enterprise clients. Coverage includes 7,000+ suburbs, 150+ metrics and 537 LGA dashboards. The main platform sits at htag.com.au. The Developer Portal offers 104+ REST endpoints and 70+ public MCP tools across six MCP servers.

The launch removes a retrieval barrier for commercial-property teams. It also gives AI developers a paid data feed. It does not solve the underlying market. Most sampled comparable sales still have no recorded price. That distinction will matter as lenders, valuers, agents and proptech teams test whether these sparse records will prove complete enough for their day-to-day commercial work.