Teams, Work and Enterprise

A team shares pools and VMs, with member, admin and billing-owner roles. Enterprise is “the same platform with limits set by contract”, on dedicated hardware or your own, with SSO, Slack support and an uptime SLA.

Sources: https://exe.dev/docs/teams/overview, https://exe.dev/docs/teams/members, https://exe.dev/docs/teams/vms, https://exe.dev/docs/teams/sso, https://exe.dev/docs/teams/sharing-controls, https://exe.dev/docs/enterprise, https://exe.dev/docs/enterprise-single-tenant, https://exe.dev/docs/enterprise-byoc, https://exe.dev/docs/enterprise-sla.

Roles

“A team always has exactly one billing owner. There can be multiple admins.” The pricing page calls the roles Member, Admin and Owner (https://exe.dev/pricing).

Creating a team and adding members

• How to create one. Teams are created through billing checkout, either by choosing Work at https://exe.dev/subscribe or by running billing manage over SSH.
• What happens to your VMs. Your existing VMs become team VMs.
• Personal subscriptions. An existing personal subscription is cancelled. The billing docs say this happens “immediately with no credit”, and you can email support about a refund (https://exe.dev/docs/billing/subscriptions).
• Inviting people. Run team add alice@example.com. Existing users have to accept, and their VMs become team VMs. New users join through the signup link. Invites expire after 24 hours.
• Listing members. team members, or the alias team ls.

Admin powers

• team vm ls lists every member’s VMs, grouped by --group=user|access|tag|region. The same list is at https://exe.dev/team/vms.
• Admins can SSH into, rename, delete and copy any member’s VM.
• team usage shows usage to admins and billing owners.
• team settings core-sharing off gives each VM a physical core of its own. Use it “if VMs may be created by different users who don’t necessarily trust each other.”
• team settings session-length 1d shortens login sessions.

SSO

By default people sign in with email and a passkey. Teams can change that:

• Google: team auth set google.
• OIDC (Okta, Azure AD and others): available on Team, Work and Enterprise.

team auth set oidc --issuer-url=https://your-org.okta.com \
  --client-id=... --client-secret=... --display-name="Acme SSO"
# set your IdP's redirect URI to https://exe.dev/oauth/oidc/callback

“Signing in with Google does not bypass OIDC.”

Sharing controls

team settings vm-sharing <off|admins-only|all-members> has three modes:

• all-members is the default.
• admins-only stops regular members from making VMs public or granting access outside the team. Sharing with the team itself still works.
• off blocks every new share for everyone, including admins. It isn’t retroactive, and removal commands keep working.

Work plan at a glance

These are the facts that matter in a deal. The full numbers are on Economics.

• Price. $75/month for the whole team, including 2 vCPU and 4 GiB. Extra capacity costs $0.105 per 2 vCPU/4 GiB unit-hour (https://exe.dev/docs/billing/work).
• Introductory price. A new team’s first Work subscription costs $37.50/month for 3 months.
• Users, pools and regions. Unlimited users, and unlimited pools in any region. Each pool holds up to 1,000 VMs.
• Allowances per pool. 100 GB of disk per 2 vCPU and 50 GB of bandwidth per vCPU (https://exe.dev/docs/pools).

Enterprise

Single-tenant and BYOC mechanics

On dedicated hosts, pool new takes no --region, because “the host decides”. pool hosts lists your hosts with each one’s region, status and the largest pool it can take. Pools “never fall back to shared hardware”.

The single-tenant page gives four reasons to choose it:

1. Defense in depth behind KVM
2. Physical isolation for auditors
3. No noisy neighbors
4. Capacity on hand without approval

The BYOC page adds three more:

1. Keeping workloads in your own account
2. Running next to your data
3. Spending existing cloud commitments and choosing your regions

SLA

exe.dev commits to “99.9% availability each calendar month, measured by exe.dev” (https://exe.dev/docs/enterprise-sla).

How credits work:

• Request credits within 30 days after month-end.
• Credits apply to future bills, aren’t cash, and are “the sole remedy for missed uptime.”
• Scheduled maintenance announced at least 24 hours ahead doesn’t count against uptime.