Basic supports one everyday assistant. Pro keeps work, home, or project assistants separate and unlocks GPU compute for custom voice and compatible open-weight models.
A capable everyday assistant on one private cloud computer.
Separate assistants, add custom voice, or run compatible open-weight models on your GPU.
Shared billing and controls for organizations using several assistants.
A recommended cloud provider, location, and computer are preselected. Review or change them before launch. No command line or cloud console required.
Cancel renewal anytime. To close everything, the “Cancel and delete account” option terminates every assistant when your paid period ends, preventing further assistant compute and storage charges.
Pro is useful when one assistant should not carry every context or connection.
Give each assistant independent memory, connected services, and settings.
Let a separate assistant handle a longer project without replacing the first.
Deploy Voice Studio or run compatible open-weight models such as Gemma on supported GPU hardware.
A separate cloud computer keeps each assistant's memory, tools, and connections isolated. Basic supports one active assistant. Pro supports several or a GPU assistant, and every running assistant creates its own compute bill.
Yes. Add supported model credentials and pay the provider directly, or let Pneum.ai handle model billing. Basic adds a 1% service charge to managed model usage; Pro includes managed billing without that added fee.
On Pro, choosing a GPU VM automatically deploys Voice Studio. You can also run Voice Studio elsewhere and connect it to any assistant through Tailscale.
Alongside hosted Voice Studio, Pro GPU agents can run compatible open-weight models such as Gemma. The model sizes and speeds available depend on the selected GPU, its memory, provider capacity, and region. Local inference uses the agent's GPU rather than managed-model requests, while compute and storage charges apply.
The Android app is currently in internal release. Paid Basic and Pro customers will be invited to the alpha when downloads open, and we will email account holders when it is ready. The browser experience is available in the meantime.
Yes. Pneum.ai preselects a recommended provider, location, and VM, then shows the estimate before deployment. You can keep the defaults or change them.
Yes. Cancel renewal and keep access through the current paid period. If you want to close everything, the “Cancel and delete account” option also terminates every assistant when the paid period ends, preventing further assistant compute and storage charges.
Yes. When you use managed model billing, set a monthly spending cap in billing settings and review how much remains before the cap resets. Optional automatic shutdown at a compute budget is coming soon.
