Install Engineering With AI.

1. Check what you need

Use Node.js 22.5 or newer, npm and Git, with one of the supported coding assistants installed. Check your versions in a terminal:

node --version
npm --version
git --version

2. Install the harness

This installs the command globally and sets up skills for the coding assistants it detects:

npm install --global @thebackstoryis/engineering-with-ai@latest
ewai install --host auto

To choose one assistant instead, replace auto with codex, claude or antigravity.

3. Open it in your project

In your terminal, go to the folder containing the project you want to work on, then run:

ewai

Follow the launcher to set up or open the project. For a project that already has code, EWAI offers Archaeology to help recover missing project knowledge. It’s optional, and you’ll be asked whether you want to run it.

4. Check the setup

Once your project is initialised, check its configuration and current state:

ewai doctor --project .
ewai checkin --project .

Read any problems reported before proceeding. Keep existing project work in version control and don’t use --force unless you’ve deliberately decided to overwrite it.

Choose the perspectives your project needs

The harness includes its core personas, but you aren’t limited to them. You can create personas inside your own project for the customers, specialists, reviewers or domain knowledge that matter to your work.

If you already have a licence, choose Set up premium personas in the dashboard and enter your key in the private form. It verifies the key and installs the pack immediately. Wait for the ready message before starting work you want the personas to help with, including Archaeology.

Your key belongs in the setup form, not in chat or a project file.

Need a different setup?

The full installation guide covers project-local installs, multiple repositories, updates and persona access. If something doesn’t work, email us for installation help with the command and error, but no passwords or keys.