Realtyna's WPL stores properties in its own table (typically wp_wpl_properties) outside wp_posts, with fields for ListPrice, BedroomsTotal, BathroomsTotalInteger, LivingArea, ListingId, MLSStatus, and dozens more depending on the field schema you enabled. Agents live in wp_wpl_users with their own MLS-aware fields. The MLS On The Fly module fetches RESO Web API feeds and writes them into the same property table, so MLS-sourced and manually-added listings share one structure.

SleekAI maps the WPL property table, the agent table, and the listing-type and location dimensions into the chatbot's system message. A buyer asking "any 3-bed homes under $600k in Glendale with a pool" gets a filtered list pulled from real rows, including agent contact info and OpenHouse dates if scheduled. The MLSStatus filter excludes pending and sold listings unless the buyer explicitly asks.

Generic chatbots see WPL's custom table and find nothing because it lives outside wp_posts. They invent properties, miss agent attribution, and can't filter by MLS-specific fields like ListingId or PropertySubType. SleekAI's data mapper supports arbitrary tables as first-class sources, so WPL's storage is read live without a sync layer.

WPL is the most-installed dedicated real estate plugin for WordPress, and the only one with a serious RESO Web API path via MLS On The Fly. The downside for AI integrations is that WPL stores properties in its own tables outside wp_posts, which means almost every generic chatbot is blind to the catalog. SleekAI's data mapper supports arbitrary tables as first-class sources, so wpl_properties, wpl_users, OpenHouse rows, and MLS On The Fly feed data all become variables the bot can query in conversation.

Buyers ask in plain language for 3-bed homes under $600k with a pool in Glendale, and the bot returns real matches filtered on ListPrice, BedroomsTotal, MLSStatus, and the amenity fields you enabled. Agent attribution flows automatically, so the right contact appears next to every listing. Open houses get surfaced with timezone-aware dates.

MLS-sourced listings stay fresh because the bot reads what the feed wrote, no sync layer needed. Display conditions plus multibot let one bot run on the buyer-facing archive, a different one on each agent profile page tuned to that agent's expertise, and a third on the broker dashboard with access to private notes. Token usage stays low because most buyer questions are short factual reads.

Bring-your-own-key keeps the brokerage in control of model and spend, which matters when listing volume gets into the thousands.