Installation

Engrams is a standalone Rust binary. It runs entirely locally in your workspace with zero daemon or server overhead. Install it once and every session your AI runs starts compounding into advice.

Step 1 — Install the CLI

Choose the installation method that fits your operating system:

Via Installer Script (macOS & Linux)

curl -fsSL https://engrams.sh/install | bash

Via Homebrew (macOS)

Install the CLI using Homebrew by tapping the official repository:

brew install stevebrownlee/engrams/engrams

From Source (Rust Toolchain Required)

If you have Rust installed, you can compile and install directly from source:

cargo install --git https://github.com/stevebrownlee/engrams-cli.git

Step 2 — Initialize your project

Navigate to your project root directory and run:

engrams init

This command initializes the local SQLite database at engrams/context.db and outputs the resolved path.

The loop starts here From the first session, your agent logs decisions and patterns as it works and recalls them on the next one. On day one your advisor mostly fetches what you ask for; within weeks the graph is dense enough that it starts anticipating — surfacing the decision that contradicts the approach it was about to take. That compounding is the product.

Step 3 — Configure your AI Tool

To allow your AI assistant to read from and write to Engrams, you must add instructions to its system prompt. See the detailed AI Tool Setup guide to set up Cursor, Zoo Code, Antigravity, Zed, or Claude Code.

Verify the installation

Run a test command in your terminal to verify Engrams is working:

engrams product-context get

It should output a default empty JSON or database path indicating it was resolved correctly.

Where is the database stored? By default, engrams looks for the closest workspace root (indicated by .git, Cargo.toml, or package.json) and stores the database at <workspace>/engrams/context.db. You can override this by passing the --db <PATH> or --workspace <PATH> flags.

Next: Quick Start →