PropETL started life answering one question extremely well: how do I get this spreadsheet into Yardi? Today it learns to answer a much bigger family of questions: what's actually in my Yardi database, and how do I work with it?
We're releasing Yardi Schema Querying — a new set of skills in the PropETL connector that puts a queryable knowledge base of the entire Voyager schema at Claude's fingertips.
What just shipped
The connector now carries, and can search, a structured map of standard Yardi Voyager:
- 9,793 standard tables — the full base-table inventory
- 137,000+ columns with types, primary/foreign key roles, and prefix conventions
- 12,900+ relationships, both inferred from Voyager's H-prefix naming convention and confirmed foreign keys
- Reusable ySQL query patterns — vetted templates with placeholders for the questions that come up on every engagement
- Join-path discovery — ask how two tables connect and get the actual join chain, not a guess
On top of the raw inventory sits a curated semantic layer: the things experienced Yardi people carry in their heads. The H-prefix naming grammar. The HMYPERSON subtype trap. Polymorphic hRecord/hCode foreign keys. Core spine joins between properties, units, leases, and people. Enum decodes for the status fields everyone has to look up. ySQL's real-world constraints.
What it feels like in practice
Inside Claude, with the PropETL connector active, you can now ask things like:
"Which tables hold commercial lease charges, and how do they relate to the lease?"
"How do I join Units to their Owners? Give me the join path."
"Draft a ySQL query listing active residential leases with their current rent, by property."
Claude consults the schema knowledge base — tables, columns, relationships, conventions — and answers with specifics: the right tables, the correct join chain, a draft query built from a vetted pattern. No more spelunking through a 60,000-line schema reference or pinging the one person who remembers how CommCharges hangs together.
This pairs naturally with what PropETL already does: the same knowledge that validates your ETL imports now explains the schema those imports land in.
Your database, not just the standard one
The standard knowledge base covers out-of-box Voyager. Real deployments aren't out-of-box — they accumulate custom tables, views, and stored procedures that no generic reference will ever describe.
That's why we also offer AI querying set up against your actual database as a custom enhancement. We index your real schema — custom tables, your view definitions, your stored procedures — into your own private overlay, physically isolated from the standard set and from every other client. Your consultants and admins then query your Yardi, with all its history and customizations, the same way: in plain English, inside Claude. Email support@propetl.com to talk about it.
If your Voyager instance has a decade of custom validators and reporting views bolted on (whose doesn't?), this is the difference between a nice demo and a tool your team actually uses.
Available now
Yardi Schema Querying is live in the PropETL connector today — every tier, including the free trial:
- Start the free 24-hour trial — full access, no credit card, license issued instantly
- Solo — Lifetime — one-time payment, one seat, permanent access, new capabilities like this one included as they ship
- Team — Lifetime — five lifetime seats on your company domain, each with its own license key, revocable and reassignable from your dashboard
Upload a messy spreadsheet, get a clean import — and now, ask the database anything on the way.