> ## Documentation Index
> Fetch the complete documentation index at: https://docs.collivo.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI permissions, context, and privacy

> Understand what Collivo AI can use and how to control its working context.

Collivo AI works with project context to answer questions and carry out supported tasks. You control access through sharing permissions and guide each task through the working folder, selected context, and AI visibility controls.

## AI works within your permissions

AI acts on your behalf and works within the permissions assigned to you. Asking a question does not give you access to a project, file, or folder that has not been shared with you.

Permissions and AI visibility solve different problems:

* **Permissions** control whether a person can open and use an item.
* **Context** focuses the agent on material relevant to the current request.
* **Hide from AI** removes an item from AI discovery tools without changing the item's sharing permissions.

Use sharing permissions when content must not be available to someone. Do not use **Hide from AI** as a replacement for sharing controls.

## Hide an item from AI discovery

1. Open the project and select the **Files** tab <Icon icon="files" color="#1F7A5A" />.
2. Find the file or folder you want to hide.
3. Open its **More options** menu <Icon icon="ellipsis" color="#1F7A5A" />.
4. Select **Hide from AI** <Icon icon="eye-off" color="#1F7A5A" />.

This control helps keep agents from discovering material that is irrelevant, unfinished, or not authoritative for routine tasks.

## Focus a task with context

Use the current working folder and the chat **Context** controls to focus an agent on the material needed for a request. Adding a file to context draws attention to that item; it does not change anyone's permission to open it.

When an answer misses expected information, check:

1. You can open the source item yourself.
2. You are working in the intended project and folder.
3. The relevant item is available to the task through the working folder or selected context.
4. The item has not been hidden from AI discovery unintentionally.

## How project data supports a request

Collivo uses the context available to a request—such as the conversation and the project material provided to the task—to produce an answer or perform an approved action.

For formal information about privacy terms and data handling, see Collivo's [Privacy Policy](https://www.collivo.com/privacy). This product guide does not replace those terms.

Avoid entering passwords, private keys, access tokens, or recovery codes in chats or files. Follow your organization's requirements when working with sensitive personal, financial, health, or client information.

## Review AI output

AI-generated output can be incomplete, outdated, or incorrect. Review important results before sharing, publishing, or relying on them for legal, security, financial, compliance, or customer-facing decisions.

Some supported actions require your approval before they run. See [Tool approvals](/ai-and-agents/tool-approvals) for how to review those requests.

## Related articles

* [Sharing and permissions](/sharing-and-security/sharing-permissions)
* [Manage access](/sharing-and-security/organization-admin-security)
* [Guide the AI with annotations](/ai-and-agents/annotations)
* [Tool approvals](/ai-and-agents/tool-approvals)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.