Table history
Review what changed on the table you have open. Table history is the right rail beside the grid — the same commit feed as Version control, filtered to one table.
Toolbar History opens the table history rail. Commits are scoped to the open table only.
Overview
Every meaningful change on a table writes one bounded commit: a CSV import, an enrichment batch, a new field, a view rename, and similar jobs. Table history lists those entries newest first without leaving the grid.
The workspace Version control tab shows the same feed across every table. Use table history when you are editing one dataset and want context beside the rows.
Open table history
- Open a table.
- In the toolbar right cluster, click History (clock icon — left of Run).
- The rail opens on the right with commits for this table only.
- Click History again or the rail close control to dismiss it.
The status bar can show the current table version short ref when a commit is selected.
What appears in the rail
Each row is one commit:
- Past-tense title — for example "Imported 847 leads" or "540 leads were enriched"
- Timestamp — absolute local time
- Actor and channel — Local user on Desktop UI, agent-as-user, API, or MCP
- Short ref — workspace-unique Crockford id from Rust (never invented in the client)
Enrichment commits show the provider logo and capability-style title. Schema adds and deletes use green or red accent strokes on the row.
Bulk work is one commit per approved job, not one commit per row or cell.
What gets recorded
Table history follows the same rules as Version control. Imports, enrichment runs, schema changes, and view create/rename/delete appear. Filter, sort, column width, show/hide, and freeze do not.
Tips
- After Import, open history to confirm one summary commit for the whole file.
- Click a table name in the workspace Version control feed to open that table with history already visible (when wired in product).
- Restore and fork from a version are planned; history is read-only navigation for now.