# Plans

## Plans in early access

Vistazo has two plans in early access, Individual for one person and Team for a group. Individual is
your personal workspace, and Team is a shared workspace that you and the people you invite build in. Both are free
while Vistazo is in early access, with no card and no trial period. The people you share with need no
plan of their own to read what you share ([How sharing works](/docs/sharing/overview)).

Single sign-on and automated provisioning (SCIM) are not part of early access. If your organization
needs them, [get in touch](mailto:support@vistazo.ai).

## The two plans

| | Individual | Team |
|---|---|---|
| Workspace | Your own private workspace | One shared workspace for up to 10 people |
| Who builds | You | Every member |
| Dashboards and notebooks | Unlimited | Unlimited |
| Storage for uploaded files ([Usage](/docs/billing/usage#how-much-can-you-own-and-store)) | 1 GB | 1 GB for each member |
| Inference and compute | A monthly allowance of each | A monthly allowance of each, for each member |
| Sharing by email or public link | Unlimited | Unlimited |
| Your own API key | Yes | Yes |

These are the standard early-access figures. Some workspaces are set up with larger allowances, more
people or no limit at all. **Settings → Usage** always shows the allowances and storage that apply to
you, and on Team the footer of **Settings → Members** shows the workspace's limit on people.

On Team you assign no license when you invite someone. Every member can create dashboards, edit the
ones they have edit access to, and run refreshes. Guests you share a dashboard, notebook or
skill with do not count toward the ten ([Members](/docs/workspace/members)).

## Inference and compute

Inference is anything that calls a model, such as the built-in assistant or repairing a chart. Compute
is code Vistazo runs for you, which today means notebook cells. **Settings → Usage** shows each as a
percentage used, and both reset each month on the day the workspace was created.

Refreshing a dashboard counts toward neither. When one allowance runs out, only the work that draws on
it pauses until the next month starts, and your dashboards keep refreshing. [Usage](/docs/billing/usage)
lists what counts toward each.

## Your own API key

An admin can add a key from Anthropic, OpenAI, Google, Fireworks AI or DeepInfra under **Settings →
Models**. The built-in assistant then runs on your provider account, and its model calls bill to that
account instead of your inference allowance. A key is also how the assistant keeps working after your
monthly inference runs out.

Notebook runs still count toward your compute allowance with a key, because no model is involved.