Skip to content

Quick start

Work through the seven steps in order. The first four get you a working system and the one idea everything else rests on. The last two are where InferNode stops resembling anything else you have run.

Everything happens on your own machine. There is no account, no cloud service, and nothing to sign up for.

StepYou end with
1Install and launchInferNode running on your machine
2Connect a modelVeltro able to answer you
3Run the tourAn agent demonstrating the system, using the system
4Everything is a fileThe idea that makes the rest obvious
5Give Veltro a taskYour own instruction, carried out
6Namespaces containAn agent failing to escape, and you knowing why
7Your work survivesKnowing what is durable and how to get yesterday back

Steps 1–4 are the short path. If you only have twenty minutes, stop after step 4 — you will have seen what InferNode is. Steps 5 to 7 are where you find out whether you want it.

  • A macOS (Apple Silicon), Linux (x86_64 or ARM64), or Windows (x86_64) machine
  • About 30 MB of RAM and 10 MB of disk for InferNode itself
  • No toolchain, no compiler, no dependencies — releases ship as signed binaries

InferNode is an operating system, not a library. It runs as a normal process on your host — a hosted Inferno — with its own filesystem namespace, its own process table, and its own shell. Nothing it does inside that namespace touches your host unless you granted it a path, and step 5 is where you prove that to yourself.