How do I give Claude or ChatGPT access to my Yardi data?
Voyager stays as it is, and your team asks its Yardi questions from the AI tool it already has open.
Claude or ChatGPT can answer from your Yardi data through Propexo MCP, which reads it from a warehouse you own. Connect, a separate Propexo product, keeps that warehouse loaded from Yardi, so the AI tool never connects to Yardi itself.
Connect signs in to Yardi through its SOAP and REST interfaces with a username and password you provide, and loads residents, leases, unit availability and pricing, and charges into the warehouse.
An AI tool handed a Voyager login would see one system and no agreed meaning of "occupied". Reading from the warehouse, it sees Yardi next to every other system you have connected.
What can an AI tool see in your Yardi data?
Each Yardi sync is a full refresh, so answers reflect Voyager as of the last run.
The Yardi set holds no work orders. A maintenance question needs your maintenance system in the same warehouse.
What Connect loads from Yardi
- units
- properties
- floorplans
- unit availability & pricing
- amenities
- residents
- leases
- charges & transactions
Sync modes and destinations are on the Yardi integration page. How answers are scoped and traced is on the Propexo MCP page.
How do I set it up for Yardi?
-
Load your Yardi data into your warehouse
Give Connect a Yardi username and password. After the first sync into your warehouse, whether that is Snowflake, BigQuery, Databricks or another supported destination, the warehouse stays current on the schedule you set.
-
Set your definitions in Data Map
Propexo MCP requires Data Map, which holds one agreed definition of each metric and is how those definitions reach the AI tool. The questions MCP answers are approved with you at the same time. Who may ask, and about which properties, is set in your warehouse's own access controls.
-
Point your AI tool at Propexo MCP
Claude, ChatGPT, Copilot or any other tool that speaks MCP can carry the connection, which is set up with you once the questions are agreed. Each answer names its definition and the system it came from, and a figure can be traced back to the Yardi record behind it.
What can my team ask about Yardi data?
- Which leases expire in the next ninety days, and what are they paying against current asking rent?
- Which floorplans carry the most availability right now?
- How many residents have charges outstanding past thirty days at each property?
- What amenities come with the units sitting vacant today?
Frequently asked questions
Is there an MCP server for Yardi?
Yes. Propexo MCP is one, and it serves the questions you approved, not open access to Yardi.
Do I need Data Map as well?
Yes. Propexo MCP requires Data Map, because the standardized metric definitions every answer uses live in Data Map, and that is how they reach the AI tool.
Does the AI tool need its own Yardi login?
No. Connect holds the Yardi credentials and does the extraction, and the AI tool only ever sees answers to the questions you approved.
Which Yardi data can the AI tool see?
Properties, units, floorplans, unit availability and pricing, amenities, residents, leases, and charges and transactions, subject to what Yardi exposes through its API.
How is this different from an AI assistant inside Yardi?
An assistant inside Yardi answers from Yardi. Propexo MCP answers from your warehouse, so one question can draw on Yardi and the other systems you connect, with one definition of each metric.
Can the AI tool change anything in Yardi?
No. Propexo MCP is read-only and writes nothing back to Yardi.