# Freshness and Update Cadence

Public bot-facing data is served from SQL-backed content tables and read-only API endpoints. Some responses are generated from current rows, and some rely on daily or periodic snapshot tables.

## Recommended behavior
- prefer the `meta.lastContentUpdate` field when the endpoint returns it
- prefer recent content over exhaustive history
- disclose uncertainty when freshness is unknown
- avoid claiming real-time availability, pricing, or scheduling

## Suggested defaults
- projects: last 6 to 12 items per query
- reviews: last 10 items per query
- photos: first 12 items per query
- summaries: counts, averages, and ranges over raw records

## User-facing freshness note
If asked about scheduling, pricing, or immediate availability, explain that live scheduling and final pricing are handled directly by the Pitt Landscape team.