For people who want to chat
Free, instant, private.
No signup. No history kept across sessions. Your prompt runs and disappears.
Open the chat
AI for everyone, powered by everyone.
Free, anonymous, instant. Your prompt runs on a real GPU shared by someone in the community — no signup, no fees.
How it works
Step 1
Send a message
Open the chat, type anything. No signup, no login, no account needed.
Step 2
The mesh finds a GPU
Your request is routed to a community-run GPU somewhere on the Tailscale mesh.
Step 3
The response streams back
You see the answer appear word-by-word as the model generates it — in real time.
Live right now
Real GPUs, real people. This page is talking to the same network /chat does.
Providers online
—
Models available
—
Latest response
—
Who's it for
For people with idle GPUs
Share your GPU. Help the community.
If you have a gaming PC sitting idle most of the day, put it to work serving AI requests. ~5 minute install.
Get started as a providerFAQ
Is it really free?
Yes. There's no fee, no signup, no account. The community runs the GPUs and we cover the small gateway hosting cost ourselves.
Who runs this?
Right now: a small group of friends running their gaming PCs as providers. Anyone with a spare GPU is welcome to join — see the 'Become a provider' page.
What model is this?
Llama 3.2 (3B) via Ollama. Every provider runs the same model so responses are consistent.
Is it private?
Your prompt is not stored on this gateway. It is, however, processed on a community member's machine for the duration of the request. Treat it like any other shared service — don't send anything sensitive.