Local LLMs: run AI on your own computer with unlimited tokens
No per-token billing, no usage caps — speed depends on your computer. Install Tutu Compute Node; no setup required.
Get started in three steps
- 1
Click download — your system is picked automatically
- 2
Double-click to install; it signs in to your account on first launch
- 3
Models are prepared automatically — ready to go
What it can do
Chat, coding, images, speech and video. Once installed, the node lists what this computer can handle based on its GPU and memory.
Where you can use it
- Tutu AI assistants: your computer is recognized automatically
- Claude Code, Codex, OpenCode: copy one command
- Any OpenAI-compatible app: paste one URL and one key
No interruptions
When your computer is off or can’t handle a task, requests switch to the network’s compute reserve automatically; if the reserve runs out, you can still queue for free. Nothing to reconfigure.
Safety
- Every frame is encrypted in transit
- No open ports — your home network address stays private
- Owner first: the moment you use your computer, it steps aside
FAQ
On your own computer there is no per-token billing and no usage cap; the only limit is your computer’s speed.
A computer with a GPU is best: NVIDIA, Apple M-series or AMD. Ordinary computers can handle lighter tasks like chat and speech.
Nothing. The installer you download already carries your account; it connects on first launch and prepares models for your computer.
Requests switch to the network’s compute reserve automatically. You don’t change anything.
No. The owner always comes first: the moment you use your computer, it pauses other people’s tasks.
The model is downloaded to your computer and runs there. On the connect page, set “Whose compute” to “My computers only” and requests go only to your own machine. Calls still go through the TutuCompute address, so you need an internet connection while using it.
Related
Local LLMs: run AI on your own computer with unlimited tokens