Formatting messages and rich answers

This page covers the formatting toolbar above the message box and the rich content a Mind can put in an answer. For the message box and its options, see The message box and its settings; to send, stop or retry messages, see Ask a Mind.

Format a message

The formatting toolbar sits above the message box.

1. Select the text you want to change, or place the cursor where you want to type.
2. Click the button you need: Undo (Ctrl+Z), Redo (Ctrl+Y), Bold (Ctrl+B), Italic (Ctrl+I), Text Size, Text Color, Code Badge, Bullet List, Numbered List or Horizontal Rule. Point to a button to see its name and shortcut.
3. Text Size and Text Color apply to the text you selected. The sizes are Extra Small, Small, Normal, Large, Extra Large and 2X Large. The colours are Foreground, Primary, Secondary, Warning and Destructive. With no selection, nothing changes.

The keyboard works too: Ctrl+B for bold, Ctrl+I for italic, Ctrl+U for underline, Ctrl+Z to undo, and Ctrl+Y or Ctrl+Shift+Z to redo. A format that is active at the cursor shows as pressed in the toolbar.

Insert a code badge

A code badge is a small named chip you can drop into a message. It holds text, SQL or code.

1. Select the text you want the badge to carry, or place the cursor where the badge should go, then click Code Badge.
2. Fill in the dialog:
   • Name: the label shown on the badge. It starts from your selection, or reads Badge when nothing is selected.
   • Type: Plain Text, SQL or Code. SQL and Code open a code editor for the content; Plain Text gives a plain box.
   • Color: Primary, Secondary or Destructive.
   • Content (shown on click): what the badge holds.
3. Click Insert Code Badge. The button becomes available once the badge has a name. To leave without inserting, click Cancel.
4. Click a badge in your message to open Edit Code Badge; change what you need and click Save Changes. To remove a badge, click the x on it. Copy copies the badge content and briefly reads Copied.

Attach part of an answer to your message

You can carry a part of an answer into your next message.

1. Select the text you want, inside a message.
2. Click Attach to message, which appears above the selection. The text is added to your message.
3. To dismiss the button without attaching, click anywhere else.

Read a rich answer

Answers come back as formatted text:

• Headings, bold, italic, bullet and numbered lists, and dividers.
• Checklists, which show as read-only checkboxes.
• Links, which open in a new tab.
• Tables and code blocks, described below.
• Charts, KPI cards and diagrams. See Charts, KPIs and diagrams in chat.
• References to artifacts, described below.

Some parts, such as links and horizontal rules, follow the Mind's colour. While the answer is still being written, a small pulsing mark trails the last line; it disappears when the answer finishes. When a model shows its thinking, that text appears smaller and muted, so it stands apart from the answer.

Work with tables in an answer

A wide table scrolls sideways inside its own frame.

1. Point to the table and click Maximize. On desktop the control appears on hover; on a touch screen it is always visible.
2. The Table preview opens with the same rows in a large view, 50 rows to a page.
3. The preview takes its title from the line above the table, usually a heading or caption. A table without a title opens as Table preview.

The Maximize control does not appear while the answer is still being written, in the model's thinking, or when the table has no readable rows.

Copy code from an answer

1. Find the code block in the answer.
2. Point to the block and click Copy code. The button briefly reads Copied!.
3. Long blocks scroll inside their frame; short blocks show in full.

A code block appears exactly as the Mind wrote it, with highlighting where the chat knows the language. It is not reformatted.

Open an artifact reference

A Mind can refer to an artifact it produced. The reference shows as a chip with its type, such as Query Artifact, Report, Presentation or Slide Deck. Presentations and reports can show as a card with a small preview instead.

• Click the reference to open the artifact.
• On desktop, use Open in modal on the reference to open it in a large view instead.
• On a phone, a tap opens the full-screen view.
• If a reference cannot be read, the answer shows ⚠️ Invalid artifact format with the raw text in its place. You can still look for the item under Artifacts.

See Artifacts to find the artifacts a chat has produced.

Questions, answered

Why can I not click Insert Code Badge?

The badge needs a Name first. Type one in the dialog and the button becomes available.

How do I edit or remove a code badge?

Click the badge in your message to open Edit Code Badge, change what you need and click Save Changes. To remove it, click the x on the badge.

How can I quote part of an answer into my next message?

Select the text in the answer, then click Attach to message above the selection. The text is added to your message.

Can I tick a checklist checkbox in an answer?

No. Checklists in an answer are read-only. Clicking a box does not change it.

What is the pulsing mark at the end of an answer?

The answer is still being written. The mark disappears when the answer finishes.

Why is the Maximize control missing on a table?

It appears only for a table with readable rows. It is hidden while the answer is still being written, in the model's thinking, and when no rows can be read.

Where does the title of a table preview come from?

From the line above the table, usually a heading or a caption. If there is no suitable line, the preview opens as Table preview with no title.

Why is a column called Column 3 in the table preview?

That column has no heading in the answer, so the preview names it by position. A repeated heading gets a number too, for example Name (2).

Is code in a chat auto-formatted?

No. A code block shows exactly what the Mind wrote, with highlighting where the language is known.

Why did an artifact reference show "Invalid artifact format"?

The text was written in a form the chat cannot turn into a reference, so the raw text shows in its place. Look for the item under Artifacts.

Why does part of an answer look smaller and muted?

That is the model's thinking. It uses a smaller, muted style so it stands apart from the answer.