Skip to content

Deploy AuraTree

Terminal window
cd docker
./aura setup

It asks for your chat and voice addresses, your storage (hosted, or the bundled MinIO) and your mail server. It writes .env and a Caddyfile, with every password, secret and key made for you, and tells you which DNS records, ports and proxy setup to put in place.

Then it checks the machine: Linux on x86-64, Docker, the ports, a shared home connection, and that your names point here. With no failures, it starts the server and prints the claim code.

  • Run it again after fixing a failed check. It keeps your .env.
  • ./aura setup --check-only runs only the checks. --settings-only only writes the files.
Terminal window
cp docker/.env.minimal.example docker/.env
nano docker/.env

Fill in every empty value. For each password, secret and key, openssl rand -hex 32 makes one. Every other setting has a default; the full list, with explanations, is in docker/.env.example.

Keep the file readable by you only: chmod 600 docker/.env.

Put a reverse proxy in front. Caddy on the same machine gets and renews certificates itself: copy the Caddyfile that ./aura setup wrote (or docker/Caddyfile.example with your names) to /etc/caddy/Caddyfile and reload Caddy.

Using Cloudflare? The chat name may be proxied, but the voice name must be DNS only (grey cloud).

Terminal window
cd docker
docker compose up -d

The first start prints a one-time claim code in the API’s log:

Terminal window
docker compose logs api | grep -A 12 'ADMIN CLAIM CODE'

Open the app, connect to your chat address, create your account, and enter the code when it asks. That makes you the server’s owner. The code is printed again at every start until someone claims it.

Once, after the first start:

Terminal window
docker compose exec -e NTFY_PASSWORD="$(openssl rand -hex 24)" ntfy ntfy user add --role=admin aura-api
docker compose exec ntfy ntfy token add aura-api

Put the token (tk_...) in NTFY_ADMIN_TOKEN in .env, then run docker compose up -d. Browser notifications (Web Push) work with no setup.

The server starts without the default soundboard sounds and emoji packs. In the app, Server Settings › Emoji (or Soundboard) offers Install the default packs. Only then does the server download them, once (about 35 MB), and check them against a checksum built into the release.

Next: harden the machine and set up backups.