Skip to main content
These settings change how every chat behaves for you. They live in Settings (profile picture → Settings), split across the General and Chats tabs.
Libra mascot
What you write under “What would you like Libra to know about you to provide better responses?” is silently prepended to the system prompt of every message you send. Treat it as a system prompt: write constraints, not bio.

​
Personalization

The General tab lets you give Libra a standing brief. It applies to every conversation, on every project.
Context Libra should always have. Best for stable facts: your role, jurisdiction, languages, the kinds of clients or work you handle. The field’s own example is “I am a business lawyer specializing in intellectual property law and copyright law. I speak German and English.”
Stylistic preferences — the example given is “I want formal, detailed and very well structured legal responses.” These apply everywhere; override them inside a single chat just by asking.
Click Save to apply. Both fields are read fresh on every message, so an edit takes effect on your next message in existing chats as well as new ones.

​
Chat preferences

The Chats tab controls defaults that apply to every new conversation. Not everyone sees the same list: the two research toggles depend on your team’s country, and the import and export actions further down depend on your role.
On by default. Chats are saved in your workspace so you can find and reuse them. Turn it off and your chats become temporary — Libra doesn’t save them, which is useful for sensitive one-off questions. Switching this toggle either way returns you to a fresh chat.
Off by default. The model button sits in the composer to the right of Research and opens the tier picker with Standard, Deep Thinking and Fast — it is hidden while an Assistant or a Workflow is driving the chat. Turning this setting on adds the individual models below those tiers, so you can run a chat on a named model instead of a tier. Refresh the page after switching it on.
Shown to German teams, on by default. It lets any chat model automatically search German statutes and case law from Wolters Kluwer Online when a question calls for it, so you don’t have to select the source under Research first. It needs no extra setup and no Wolters Kluwer subscription, and it replaces the older Otto Schmidt default — Otto Schmidt stays available as a source you pick yourself. Refresh the page after toggling.
The counterpart outside Germany, for teams whose country has Wolters Kluwer access. On by default where it appears: every new chat starts with Research already switched on and your country’s Wolters Kluwer sources already selected, so you don’t have to pick them each time. Turn it off and new chats stop pre-selecting them — you choose the source yourself under Research when a question calls for it. Refresh the page after toggling.

​
Bulk chat operations

Lower down the Chats tab, two actions apply to all of your chats at once. Both replace themselves with an Are you sure? confirmation before anything happens. Administrators get two more actions on the same tab: Import Chats and Export Chats. Export downloads all of your chats as a single JSON file, useful for backups or migration; Import reads such a file back in. They appear only for the accounts that also see the Admin Settings tab, so most people won’t find them.
For deleting or archiving a single chat, use the chat menu instead — see Chat history.

​
Next steps

Chat modes

Standard, Deep Thinking and Fast — pick the reasoning depth for the message you send next.

Settings

Profile, security, templates, and integrations.