Documentation

Introduction

Engram is a local-first AI operating system designed for high-liability environments. It runs completely offline, turning your machine into a private intelligence hub.

Quickstart Guide

Get Engram running on macOS or Linux in under 5 minutes.

Architecture

Learn how the local vector database and inference engine work.

Why Engram?

Most AI tools operate in the cloud, sending your sensitive logic, code, and personal data to remote servers. Engram inverts this model.

Local-First: Weights, biases, and vectors never leave your device.
Air-Gapped Ready: Once installed, zero internet connection is required.
High Performance: Native optimization for Apple Silicon (Metal) and NVIDIA GPUs (CUDA for Linux & Windows).

System Requirements

RAM16GB Unified (Min)

Storage20GB SSD

OSmacOS 14+ / Windows 10+ (WSL2) / Linux