Does iDBQuery cache query results, and how fresh are the numbers?
For live databases and warehouses, iDBQuery runs a fresh query each time, so numbers reflect the current state of your data. File, document, and API sources are worked from their last synced copy, which you refresh on demand or on a schedule. Every answer cites the source it read from.
Freshness depends on the type of source, and iDBQuery is explicit about it.
- Live sources. For connected databases, warehouses, and read replicas, iDBQuery queries in place at the moment you ask. There's no pre-built extract sitting in between, so a figure reflects the data as it is right now.
- Ingested sources. Files, folders of documents (via OCR), and REST APIs are brought in and indexed. iDBQuery answers from that synced copy, and you control how often it refreshes — manually, or on a schedule you set per source.
- Consistency within a conversation. Follow-up questions build on the same context so a thread of related answers stays coherent.
This design gives you live accuracy where it matters and predictable refreshes where a source can't be queried live. Because every answer names and cites the source it read, you can always tell whether a number came straight from the live database or from the last sync of an uploaded file — and refresh that source if you need the very latest before making a decision.
Updated 2026-06-22