Back to directory

NightRun

A Rust, UEFI-resident LLM runtime that boots a PC or Raspberry Pi 5 directly into a local chat prompt from a USB stick or SD card, with no operating system underneath.

NightRun

NightRun is a local LLM runtime that boots straight into a chat prompt: the machine's firmware starts it directly from a USB stick or SD card, with no Linux userspace, kernel, browser, network stack or host process underneath. It streams a quantized model, 1.3 to 2.4 GB, into RAM with checksums verified during the read, then seals storage so any later disk read is a deliberate hard fault.

It runs on ordinary x86_64 PCs and on a Raspberry Pi 5, drawing its own terminal on the framebuffer with a custom renderer and bitmap fonts, reading USB keyboard input, and running inference across multiple cores through firmware services; on the Pi 5 it also drives the cooling fan directly, since nothing else is present to do so.

An installer builds a bootable image and flashes it to removable media after a verified, checksummed model download, refusing to target a system disk and requiring the exact device name typed back before writing. Building and testing it requires Linux, Rust and, for testing without hardware, QEMU.

More in Inference

See all 770
  • OpenDecision

    deepanwadhwa.github.io

    An open-source Python package that answers typed questions about application state or documents using a local natural-language-inference model, returning structured choice, score, yes/no or evidence-relation values instead of free text.

  • Lifeboat

    iterate.ai

    A local inference server that puts an OpenAI- and Anthropic-compatible API in front of open-weight language models running on hardware you control, from a laptop to a GPU cluster.

  • Claude S40

    github.com

    An unofficial Claude chat client for 2007-era Nokia Series 40 phones, built as a Java ME app plus a small Go relay server, with web search, reading mode and saved chats.