Notebook Workspaces

A workspace is a shared, persistent file system that you open inside JupyterLab. It holds the notebooks you and your team work on, the SSH keys you need for source control, and any data files you upload while exploring. The workspace lives between sessions — close your browser, come back tomorrow, and your files, kernels, and JupyterLab tab layout are right where you left them.

Workspaces are also the unit of access control on the platform. Everything inside a workspace is visible to every member; nothing inside it is visible to anyone outside. Membership is the boundary.

Group workspaces and personal workspaces

There are two kinds of workspace:

• Group workspaces are created by an admin for a team and shared with that team. Use these for any work that more than one person will touch — team analyses, shared notebook cards, dashboards in development.
• Personal workspaces are created automatically for every user. You don't have to ask an admin to set one up — when your account is provisioned, a workspace just for you appears in your workspace list. Use it for scratch work, one-off explorations, and anything you don't want cluttering a shared space.

Requesting a group workspace

Group workspaces are created by your platform admin. To get one for your team, contact your admin and tell them: the workspace name you want, and which user accounts should be members. Once created, the new workspace shows up on your workspace landing page automatically. Membership changes (adding or removing teammates later) are also done by the admin — there is no self-service member management today.

Sharing a single notebook

There is no per-notebook share control. The unit of sharing is the workspace. To share one notebook with a teammate who isn't already in the workspace it lives in, the simplest path is to copy the notebook into a workspace you both belong to:

• In your source workspace, download the notebook from the JupyterLab file browser (right-click → Download).
• Open the destination workspace and upload the notebook there.

If you don't share a workspace yet, ask your admin to either add the teammate to your existing workspace or create a new group workspace with both of you as members.

Opening a workspace

Workspaces are listed on the workspace landing page in the Syntasa portal. Each entry shows the workspace name, members, and a launch action. Clicking launch opens JupyterLab in a new browser tab; the file browser on the left shows the workspace's flat root directory, and any tabs you had open last time are restored.

The workspace layout

When you open a workspace in JupyterLab, the file browser shows a single flat root. Everything in the workspace lives at the top level — notebooks, team folders, configuration, and any data files you upload.

The post-9.1 workspace layout. All workspace members see the same flat root.

Three folders are present in every workspace:

• notebooks/ — every notebook card (a notebook tracked by the portal — see Notebook Cards vs JupyterLab Notebooks) created in this workspace lands here. Cards from every workspace member are merged into this single folder; there is no per-user subdirectory and no ownership marker on a card folder. Name your cards meaningfully — that name is what other workspace members will see.
• config/ — system-managed. The platform stores the workspace's SSH keys here so you can connect to source control from inside a notebook. Don't edit anything in this folder by hand.
• Any other folder at the root — these are folders you or another workspace member created. Use them to organize JupyterLab notebooks, datasets, and scripts the way that makes sense for your team.

On workspaces that were upgraded from 9.0.x, you will also see a users/ folder at the root, preserved by the migration and described under Upgrading from 9.0.x below. Workspaces created fresh in 9.1 do not have a users/ folder at all — there is no per-user private area inside a workspace. For private work, use your personal workspace instead.

Sharing within a workspace

Within a workspace, nothing is private. Every member can see every file, open every notebook card, and read into any folder under the workspace root. This is intentional — workspaces are designed for collaboration, and the platform's storage layout reflects that. The right way to control who can read your work is to control who is a member of the workspace it lives in.

Upgrading from 9.0.x

If you used Syntasa Notebooks in 9.0.x, the workspace layout looked different. Workspaces were divided into a public shared/ folder and per-user users/{username}/ folders. The shared/ folder was the only collaboration surface; everything under users/ was visible only to the user who owned it. Notebook cards were stored under each user's private folder, so you could not see another team member's cards from JupyterLab.

Figure 3. The pre-9.1 (legacy) workspace layout: dual-zone hierarchy with shared/ and per-user private folders.

9.1 collapses that two-tier model into the single flat root described above. The first time you open an existing workspace after the upgrade, an automatic migration runs once and rearranges your files. You don't have to start it; nothing in your workflow needs to change.

Here is what the migration does:

Three details about the migration are worth flagging:

• Nothing is deleted. The original shared/ folder is preserved as a hidden backup at the workspace root. If you need to compare against the pre-upgrade state, the data is still there.
• The users/ folder stays. Files under your old users/{you}/ private area are not moved. They remain at users/{you}/ — but they are now visible to every workspace member. Move anything you want to keep private to your personal workspace.
• Saved notebook processes still work. If you have scheduled notebook processes that point at the old shared/ path, the platform translates the path automatically when the process runs. You don't need to update them.