Bring your compute together.
Share the computing power your Ameeba Workspace has available. Run more Max work through a Compute Pool or connect compatible devices in a Compute Cluster.
Ameeba Compute
Available power becomes shared capability.
Turn on Share Compute, choose how much your Ameeba Workspace can offer, then decide which chats can use it. Ameeba handles the rest.
Shared compute for this workspace
Available power, under your control.
Share Compute
Your Ameeba WorkspaceYour activity always takes priority.
75%10% 18% available now100%
Compute PoolMore Max work at once.
Independent Pseudopods take independent jobs as capacity becomes available.
Compute ClusterRun a larger Max model.
Compatible Pseudopods can cooperate on one larger model when the cluster is ready.
Pseudopods3 online
You · AmeebaNVIDIA GeForce RTX 5060 Ti · 15.9 GB VRAM
Up to 75% available18% available now
SecOps · Ameeba WorkspaceNVIDIA GeForce RTX 4070 · 12 GB VRAM
Up to 50% availableAvailable
ReleaseOps · Ameeba WorkspaceNVIDIA GeForce RTX 4080 · 16 GB VRAM
Up to 60% availableAvailable
Chat scoped. Only members of this workspace can use compute shared here.
How it works
Three controls. One simple model.
Your Ameeba Workspace decides how much can be contributed. Your chat decides who can use it. Ameeba Compute decides how available capacity is scheduled.
Your Ameeba Workspace
Turn on Share Compute and set the maximum percentage Ameeba may use when capacity is available.
Your chat
Enable Share Compute only in the chats where you want your contributed capacity to be available.
Ameeba Compute
Available capacity is scheduled across eligible work without exceeding the limit set on the device.
Two ways to compute
Pool for more work. Cluster for larger models.
The same contributed Pseudopods can serve different kinds of workloads without pretending separate GPUs are one device.
More Max work at once.
Independent Pseudopods run independent jobs. More available workers increase total throughput across the chat.
JobAmeeba WorkspaceJobAmeeba WorkspaceJobAmeeba Workspace
Run a larger Max model.
Compatible Pseudopods cooperate on one model or inference workload. Model, runtime, network, and hardware compatibility determine what can cluster.
8 GB12 GB16 GB
One coordinated workloadControl stays local
Share what is available. Keep what you need.
Set a maximum once on each Ameeba Workspace. Ameeba can use capacity up to that limit when it is actually available, while your own work continues to take priority.
50% means up to 50%. It does not reserve half of your machine.
Built around consent
You decide what joins.
Share by choice
Share Compute is opt in. Joining a chat never contributes your Ameeba Workspace automatically.
Set your limit
Choose how much compute Ameeba may use. The limit is a ceiling, not a reservation.
Your activity comes first
Ameeba uses capacity when it is available and backs off as your own workload needs the machine.
Choose each chat
Device sharing and chat permission both have to be on before a chat can use your compute.
Security
Know the trust boundary.
Encrypted transport protects data while it moves. A Ameeba Workspace processing remote work still handles that workload. Share Compute only with chats and devices you trust.
Open Security CenterPut the compute you already have to work.
Install Ameeba Workspace, turn on Share Compute, and choose where your available capacity can help.
Get Ameeba Workspace