FacetWP indexes each post's facet values into the wp_facetwp_index table, with one row per facet-name plus post ID plus value plus display label combination. SleekAI's Wizard maps that table by facet name, so the bot can ask which values exist for a facet, which posts match a partial set of facets, and how many results a given combination yields without re-running FacetWP's query engine itself.

Yes. The Wizard exposes a count function that joins facetwp_index against the active query, so the system prompt can resolve a facet combination into an integer count. That lets the bot tell a visitor that adding pets_allowed brings the count from 14 to 6, useful for keeping users out of empty-result dead ends on busy directory or catalog sites.

The Layout Builder is a Pro feature for designing results templates. SleekAI does not interact with the Layout Builder directly. It reads the underlying facets and posts, which is what the templates ultimately render. The bot operates on data the visitor would see in the template if it were not over-filtered, so its suggestions stay aligned with the actual results template across all layouts you build.

Yes, with one caveat. Search facets and proximity facets compute on the fly against the post content and lat-lng pairs in postmeta. The Wizard maps the underlying fields so the bot understands the geographic and search dimensions, but it does not run the proximity calculation itself. For most cases the bot quoting nearby city names and approximate radii is sufficient to keep the visitor moving.

Yes. FacetWP exposes a no-results callback on each template. Call the SleekAI JS API from that callback to open the chatbot with the current facet selections as initial context. The bot then proposes specific relaxations grounded in counts from wp_facetwp_index, which converts a visible dead end into a recoverable conversation about what the visitor actually wants.

Yours. SleekAI is bring-your-own-key, so message costs are billed directly by your provider with no SleekWP markup. Use OpenAI, Anthropic, Google, or OpenRouter, and pick a fast model for routine facet-count questions and a stronger one for reasoning about complex search intent under the same chatbot without changing the integration.

No. FacetWP's refresh handler runs over its own AJAX endpoint against the facetwp_index table. SleekAI runs over a separate chat endpoint. The chat widget is lazy and only initializes when the visitor opens it, so neither initial page load nor the facet refresh round-trip is affected by SleekAI being installed and active on the same page or template.

Yes. The SleekAI JS API can read FacetWP's active selections from the FWP object and pass them into the chat request as initial context. That lets the system prompt resolve the visitor's currently applied filters, so the bot can react to whatever they have set in the moment they ask, rather than answering as if the page had no active filters at all.