Full transcript

By default K-code keeps threads compact: older turns fold, work groups collapse, thinking shows a summary and long tool output is capped. Full transcript mode turns all of that off so you can read exactly what the agent did.

What it changes

• Turn folds and work groups are expanded, without changing your own open/closed choices (turn it off and they return).
• Thinking shows in full, rendered as markdown.
• Tool calls show their full input and full output. Command output that was omitted from the live view is fetched when its row renders.
• Height caps on logs are removed.
• Agent Hub panes and logs show full thinking and full tool input and output.

Long threads render the same way, just taller.

Turn it on

• Settings → Appearance → Full transcript.
• The Toggle full transcript button in the chat header.
• The command palette action Show full transcript (Hide full transcript when on).
• The keybinding chat.fullTranscript.toggle, bound by default to mod+alt+shift+e (change it in Settings → Keybindings).
• Mobile: Settings → Appearance → Text and the button in the thread header (iOS and Android).

The setting syncs across your devices connected to the same server; see Settings sync.

Found a mistake?Edit this page on GitHub.