Settings, usage limit and safety modes
Open in Guilda WorkHow many queries you have left, keys, read-only mode and shortcuts.
Updated on
- AI assistant
They are under AI Assistant settings, reachable from the assistant's chat.
Your firm's AI usage
The top shows the queries used this month against your firm's limit. When few are left, it warns you. If it runs out, the assistant stops answering until it renews on the 1st of the month; to extend it, contact support. The limit is set by the platform administration for each firm and can be unlimited.
Model and keys
Choose a model from the list of free models or type any OpenRouter id (including paid ones). You can store several API keys, one per line and in order of preference: if the first fails for a reason specific to that key (no credit, invalid, limit reached), it switches automatically to the next. You can delete all the saved keys whenever you like.
Two safety modes
- Autonomous mode: the assistant does not ask for confirmation when creating, editing or deleting content. Even so, actually sending an email always asks for explicit confirmation, and the assistant can never delete backups.
- Read-only mode: the assistant only looks things up; it never creates, edits or deletes.
Prompt shortcuts
Shortcuts appear above the chat's text box and, with one click, send their text. Add your own with a short title and the text (maximum 20 of your own) and remove the ones you don't use.
Related guides
Using the AI Assistant
Ask questions, request summaries and let it carry out actions for you.
Connect ChatGPT or Claude to Guilda Work
Use your notes, tasks and clients from ChatGPT or Claude with a connector.
Email reminders with the assistant
Ask it to remind you of something or warn you before a deadline.
Guilda Work