Kazi AI privacy and limits
What Kazi AI sees, how your content is processed, your monthly token allowance, and the limits of AI-generated content.
What Kazi AI sees
Kazi AI only sees what you send it, when you ask: the selected text, the open file's relevant content, or your prompt. It doesn't read your other files, your mailbox or your vault in the background.
Requests are processed by Kazi's AI service, which uses a third-party large language model to generate the answer. Don't send information you're not allowed to share outside your organisation, such as other people's medical or financial records.
Allowance
| Who can use Kazi AI | Anyone with a free Kazi account |
| Monthly allowance | 10,000,000 tokens |
| Resets | On the 1st of each month |
| Remaining | Shown on the Kazi AI button in the top bar |
A token is a piece of a word: about 1,000 tokens make 750 English words. Both your request and the answer count.
Limits of AI
- AI can be confidently wrong. Check facts, figures, names and quotes.
- AI doesn't know about events after its training, or anything private you haven't given it.
- You're responsible for what you send. Review AI-written emails and documents before they go out.