MentionFlow AI
Open app
DocsFeature guides

Batch actions and search

Manage a large prompt list at scale — select many prompts at once, search and multi-tag filter to narrow the set, act on the whole selection in one click, and show or hide table columns.

When your prompt list grows past a screenful, batch actions let you work it like a spreadsheet: search and filter to the exact set you want, select it, and apply one change to every prompt at once. This page covers the power tools on the Prompts table — selection, bulk actions, search, multi-tag filtering, and column choices.

What it does

The prompt table lets you:

  • Select many prompts with checkboxes, a select-all, and shift-click ranges.
  • Act on the whole selection in one click — set a topic, add or remove tags, set a location, pause, resume, archive, or restore.
  • Search the list by prompt text or tag.
  • Filter by several tags at once, with AND or OR matching.
  • Show or hide columns so the table shows only what you care about.

Who it's for

  • Agencies and larger teams organising hundreds of prompts across topics, tags, and stages.
  • Anyone cleaning up — retagging a campaign, archiving a batch to free plan slots, or moving a set of prompts into a topic folder.

Try it: select and act in bulk

  1. Open Prompts on a live workspace. (Selection and bulk actions are hidden in the demo, and require editor access.)
  2. Tick the checkbox on any row, or the header checkbox to select every prompt currently matching your filter. Hold Shift and click a second checkbox to select the whole range between them.
  3. A toolbar appears showing "N selected". If you have a partial selection, a "Select all N in this filter" button selects every matching prompt.
  4. Pick an action from the toolbar:
    • Topic — move the selection into a topic folder, or to No topic.
    • Tags — Add tags… or Remove tags…, typing a comma-separated list.
    • Location — set a per-prompt location, or Auto (project region).
    • Pause, Resume, or Archive the selection. In the archived view the action is Restore.
  5. Read the result line. Actions never fail silently: you'll see, for example, "14 archived, 2 failed:" with the reason. Prompts that succeeded leave the selection; ones that failed stay selected so you can retry. Use Clear to drop the selection.
Note

The selection follows your filter. If you change a filter or the search box so that a selected prompt is no longer shown, it drops out of the selection and the toolbar tells you how many were removed. Archiving frees a plan slot; restoring re-checks the batch against your quota — see Plans and quotas.

Try it: search and filter

  1. Type in the "Search prompts & tags…" box at the top of the list. It matches your text against both prompt wording and tags, and updates as you type. Press Escape or the clear button to reset it. (In the archived view, search matches prompt text only.)
  2. Click one or more tag chips under Tags to filter by tag. Selecting two or more tags reveals a Match toggle: All tags keeps only prompts carrying every selected tag (AND), while Any tag keeps prompts carrying at least one (OR). The default is All tags.
  3. Combine search, tags, stage, category, intent, and topic freely — every filter is part of the page URL, so any view you build is a shareable link.

Try it: choose your columns

  1. Click the columns button (its tooltip reads "Show / hide columns").
  2. Toggle any of Type, Location, Volume, Sentiment, Mentions, Engines, Visibility, Avg position, Change (Δ), Tags, and Topic. The prompt text, checkbox, and row-actions stay fixed.
  3. Use Reset to default to show every column again.

Your column choice is remembered in your browser (it is a personal, per-device preference, not a workspace setting), so it does not change what your teammates see.

How it's computed

  • Batch size and quota. A single batch is capped at 1,000 prompts. Pause, resume, and archive are quota-neutral; restore re-checks your plan limit inside a locked transaction, so a bulk restore can never push you over your quota.
  • Which actions exist. Bulk covers topic, tags, location, and the status moves pause / resume / archive / restore. There is no bulk "set stage"; resume is the paused-to-active path and restore is the archived-to-active path.
  • Search and tag matching. Search is a case-insensitive "contains" over prompt text and tags. Tag filtering supports AND ("all") and OR ("any"), defaulting to AND; both are stored in the URL.
  • Columns. The table remembers which columns you hide, so any column added in a future release defaults to visible. Preferences live in your browser only — the app has no server-side UI-preference store. Column order is fixed; only visibility is adjustable.
  • Auto-tag honesty. Some tags are applied automatically at creation (for example the shopping-intent and branded tags). Because there is deliberately no backfill, prompts created before the auto-labeler shipped are not tagged — so filtering on an auto-tag can undercount. The filter chip's tooltip says so.

Limits

  • Live workspaces only, and editor rights required — view-only members and the demo see no selection or bulk toolbar.
  • A prompt pushed over the per-prompt tag limit fails for that row only, with the reason shown.

Related