Live row reads

SheetDB exposes each spreadsheet as a JSON endpoint, and an agent queries it to read rows on a trigger or schedule. It pulls the columns your rule needs, checks values against your criteria, and passes clean records downstream to a CRM, form, or another agent. Rows that read cleanly move on without review. When a required column is blank, a value fails a format check, or two rows describe the same entity with different details, the agent stops and routes the record to the sheet owner, who corrects the source data before it feeds another system.

Row inserts and edits

Data arriving from your other tools needs to land in the sheet, and the agent writes it through the SheetDB API. It reads the incoming payload, applies your mapping rules to place each value in the right column, and either appends a new row or updates an existing one by key. Straightforward writes that fit the schema post on their own. When a write would overwrite a filled cell, a key matches several rows, or a field falls outside allowed values, the agent holds the change and notifies a person to confirm before anything is saved.

Filtered search queries

Instead of scanning a whole sheet, an agent uses SheetDB search parameters to fetch only the rows that meet a condition, such as a status, date range, or owner. It reads those rows, applies your rule, and acts on the subset, which keeps lookups fast and predictable. Queries that return a clear result are used as is. When a search finds nothing where a match was expected, brings back far more rows than your rule anticipates, or hits an API limit, the agent flags the query and asks a person whether to widen or narrow the criteria.