> ## Documentation Index > Fetch the complete documentation index at: https://docs.youniverse.ai/llms.txt > Use this file to discover all available pages before exploring further. # Vault and logins > Give Agents passwords and keys they can use but never see. Glass safe opened by a robot hand without seeing inside The **Vault** stores logins and keys locked. The Agent can use them, but **never reads them and never repeats them back**. ## What you can save | Type | Use it for | | - | - | | **Login** | Website, username or email, password, optional notes ("use this account only for job applications") | | **Key** | An API key or token for a service | | **Key pair** | A public and private key or ID + secret | | **Note** | Any private text the Agent should be able to use | ## How secrets stay secret * Each item is encrypted with your own key. * The AI only sees the item's **name and type**, and inserts a placeholder. The real value is filled in at the last moment, outside the model. * Decrypted values are never written to run history or logs. Glass browser window with replay and folders ## Shared or private The Vault works like one shared computer: items are available to **all your Agents** unless you mark them **This agent only**. You can also pick exactly which other Agents may use a private item. ## Save a live login If you sign in yourself while watching an Agent's browser, tap **Save it to Vault** so the Agent can sign in on its own next time. ## When an Agent needs a credential It shows a compact card listing the credential types it needs, with vendor setup steps tucked under **Technical details**. Yes, you can reveal your own items with the eye icon. Agents cannot. Never. Passwords, vault items and memories are never copied. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.