Your AI.Your machines.Usable by your team.
Publish the models, tools and agents already running on your own hardware as private services your team can call. No public ports, on any machine.
demo-laptop
Three commands from laptop to team-ready.
-
01
Connect
Joins this machine to your private network.
fabric up -
02
Publish
Turns a local model, tool or API into a named private service.
fabric serve http://127.0.0.1:11434 --name my-model
Ollama
vLLM
llama.cpp
LM Studio
OpenAI API
MCP
Hugging Face
NVIDIA
Apple
Docker- and the rest of your stack
-
03
Share
Mints a scoped, expiring token for a teammate, or for a customer demo you can revoke in one click.
fabric grant llm://my-model
Nothing here is a mockup.
Terminal
Console
Your prompts never leave your machines.
Agent-Fabric cloud
sees only
nothing below crosses up
your laptop
encrypted mesh
your GPU box
We couldn't read your prompts if we wanted to.
- Signed releases and network maps
- Full audit trail
- Revoke any access instantly
Flat per workspace.
A ten-person team pays the same as a three-person team.
| Capability | Free₩0 | Builder₩39,000/mo | RecommendedTeam₩129,000/mo | Business₩399,000/mo |
|---|---|---|---|---|
| Networks · machines | 1 · 3 | 3 · 25 each | 10 · 100 each | unlimited |
| Members | 3 | 5 | 15 | unlimited |
| Customer demo rooms | 1 | 10 | unlimited | unlimited |
| GPU & agent workspaces | Not included | 3 + 3 | unlimited | unlimited |
| Customer environments | Not included | 1 | 10 | unlimited |
| Remote local-console access | Not included | Not included | yes | yes |
| Relay traffic | 5 GB/mo | 50 GB/mo | 500 GB/mo | 2 TB/mo |
| Usage history | 7 days | 30 days | 90 days | 365 days |
| Support | Community | SLA, dedicated channel | ||
| Self-hosted control plane | Not included | Not included | Not included | yes |
| Start free | Subscribe | Subscribe | Contact us |
Per workspace, per month. VAT included.