Installer Script¶

The automated installer script sets up the full ai.doo stack on a fresh Linux server in under 10 minutes.

Download and Run¶

curl -fsSL https://raw.githubusercontent.com/aidoo-systems/ollama/master/scripts/install.sh | bash

Or clone the repo and run directly:

git clone https://github.com/aidoo-systems/ollama.git
cd ollama
./scripts/install.sh

What It Does¶

Checks prerequisites (Docker, Docker Compose, curl)
GPU detection (NVIDIA via nvidia-smi)
Interactive product selection (PIKA, VERA, or both)
Secret generation and directory setup
Docker image pull and service startup
Health checks and readiness verification

Options¶

Flag	Description
`--no-gpu`	Skip GPU detection, use CPU mode
`--products pika,vera`	Skip product selection prompt
`--password <pass>`	Set admin password (skip interactive prompt)
`--with-caddy`	Add Caddy reverse proxy with automatic HTTPS
`--domain example.com`	Base domain for Caddy subdomains (requires `--with-caddy`)
`--acme-email admin@example.com`	Email for Let's Encrypt notifications
`--yes`	Accept all defaults (non-interactive)

HTTPS with Caddy¶

To deploy with automatic TLS and hostname routing:

./scripts/install.sh --with-caddy --domain example.com --acme-email admin@example.com

This generates a Caddyfile with hub.example.com, pika.example.com, and vera.example.com subdomains, binds all service ports to 127.0.0.1 (so only Caddy is publicly reachable), and configures VERA's CORS and API URLs for HTTPS.

Prerequisites: DNS A records for all three subdomains must point to the server, and ports 80/443 must be open.

Directory Structure¶

After installation, your ~/aidoo/ directory will contain:

~/aidoo/
├── .env                 # Environment variables
├── docker-compose.yml   # Stack definition
├── Caddyfile            # Reverse proxy config (if --with-caddy)
├── secrets/             # Docker secrets
│   ├── hub_admin_password
│   ├── hub_auth_api_key
│   └── hub_secret_key
└── data/                # Persistent data
    ├── ollama/          # Model storage
    ├── hub/             # Hub database
    ├── pika/            # PIKA documents + ChromaDB
    └── vera/            # VERA uploads

Next Steps¶

After installation:

Open Hub at http://your-server:2000 (or https://hub.example.com with Caddy) and log in with your admin credentials
Pull a model from the Models tab (e.g. llama3.2:3b)
Activate your license key in the License tab
Open PIKA at :8000 or VERA at :3000 and log in with your Hub credentials

For production deployments with TLS and hostname routing, see the Reverse Proxy guide.