Workspaces in Ayeto: Persistent Memory for Your AI Assistants

What Are Workspaces, and What Problem Do They Solve?

Workspaces are persistent working environments where your AI assistants can store and reuse notes, files, and supporting materials—even after a conversation ends. In other words, they give an assistant memory that does not disappear when you close the chat.

The problem is familiar. You develop a useful analysis with an assistant, define the project context, and summarize decisions from a meeting. A week later, you start from scratch. Conversation history helps you find an old exchange, but it is not a structured repository. Workspaces close that gap: your materials live in one place, can be shared across multiple assistants, and remain under your control.

A workspace can be personal or organization-based, depending on the context in which you create it. Organization Workspaces follow the roles and permissions of the relevant organization.

How a Workspace Works

Every workspace has a name and description and can contain two types of items:

• Notes — text stored directly in the workspace, such as decisions, meeting notes, briefs, or checklists.
• Attachments — copies of files stored in the workspace. If you later delete the original file from a conversation, the workspace copy remains available.

You can organize items into a simple hierarchy using a parent item (parent_id). A typical project structure might look like this:

Project overview
├── Requirements
├── Decisions
├── Meeting notes
└── Attachments

Clear titles and short descriptions are worth the effort. They improve navigation and help the assistant retrieve relevant information through semantic search.

Two Ways to Work: Chat and the User Interface

You can manage a workspace in two ways. Both operate on the same underlying content.

1. Through an AI Assistant in Chat

You do not need to click through menus. Tell the assistant what you need:

Create a workspace called “Project Alpha” with the description “Project materials, decisions, and outputs.”

Save a note called “Meeting decision” in “Project Alpha” with this content: Production deployment will take place on July 20.

Find information about the deployment date in the workspace.

The last example is particularly useful. The assistant supports semantic search across workspace content. It does not look only for exact keywords; it searches by meaning. By default, it returns up to 5 results, and the limit can be increased to 50.

2. Through the User Interface

If you prefer manual control, open Workspaces in the left-hand menu. Each workspace has four tabs:

• Content — create and edit notes, upload files, and open stored materials.
• Assistants — choose which assistants can access the workspace and with which permissions.
• Sharing — grant access to other users and organization groups.
• Settings — edit the name and description, archive the workspace, or delete it.

A workspace is also available directly in chat. When an assistant is selected, use the Assistant Workspace button to open a side panel next to the conversation. You can read supporting materials, edit notes, and upload files without leaving the chat.

Access Control: Who Can Do What?

Access is managed on two separate levels.

1. An assistant’s access to the workspace, configured under Assistants:

• Not connected — the assistant cannot see the workspace.
• Read only — the assistant can list, search, and read items.
• Read and write — the assistant can also create, update, and delete items.

2. Sharing the workspace with people, configured under Sharing. You can grant access to individual users and, for organization Workspaces, entire organization groups.

These levels are separate by design. To use a workspace through an assistant, both the user and the assistant must have access. Give Read only access to assistants that should analyze source material without changing it. Use write access only when an assistant genuinely needs to manage content.

Tip: If you connect a workspace from the chat panel using Assign workspace, the assistant receives Read and write access. If it should only read the content, change the permission afterward under Assistants in the workspace detail.

Practical Use Cases

• Project knowledge base — create one workspace per project. Store briefs, decisions, and meeting notes as separate items instead of leaving them buried in chat history.
• Shared context for multiple assistants — connect one workspace to several assistants. A coordinator can write updates while specialists have read-only access.
• Continuity for long-running work — let an assistant save intermediate findings during a multi-step task, so you can return to them a week later.
• Company materials with controlled access — share an organization Workspace with a group instead of managing permissions for each individual file.

What to Watch Out For

• Search in the UI is not semantic search. The interface filters items by title and description. Meaning-based search inside notes and documents is performed by a connected AI assistant.
• File contents cannot be edited in the UI. You can change an attachment’s title and description, but a revised file must be uploaded as a new item.
• Some fields save after you leave them. Names, descriptions, and note contents in parts of the interface are saved when you click outside the field.
• Deletion is permanent. Deleting an item or an entire workspace cannot be undone. For completed projects, use archiving instead. The content and assistant connections remain intact, while the workspace disappears from the default overview.
• Organization context matters. If you cannot find a workspace, check that you are using the correct personal or organization context and that the workspace is not archived.

Recommended First Setup

1. Open Workspaces in the main menu and verify the active context.
2. Create a workspace named after your project or topic.
3. Add the core notes and files under Content.
4. Under Assistants, connect only the assistants that need the materials and assign the appropriate permissions.
5. Under Sharing, add colleagues or organization groups.
6. Open the workspace panel in chat and verify that the assistant can access the correct materials.

Workspaces turn an assistant from “a smart chat that forgets everything” into a collaborator with persistent memory and shared context. Create your first workspace for an active project and give your assistant something useful to remember.

Try Workspaces today and keep important project materials where your assistants and your team can actually find them.