Installation Guide

InfraMind is designed for fast, portable deployment. Any Linux-based system that meets minimum requirements can be transformed into an active compute node by running a single command. This installation flow handles all provisioning, configuration, and background execution required to integrate a machine into the InfraMind mesh.

The default installation is performed via a remote shell script:

curl -sL https://inframind.host/install.sh | bash

This script performs a sequence of idempotent setup operations. Each stage is validated, and fallbacks are applied based on the host’s operating system and available resources.

Step 1: Docker Installation

The script first checks for the presence of a working Docker daemon:

if ! command -v docker &> /dev/null; then
  echo "[InfraMind] Docker not found. Installing..."
  curl -fsSL https://get.docker.com | sh
fi

The script uses the official Docker convenience installer to support a wide range of Linux distributions. For hardened environments or custom security policies, pre-installation of Docker is recommended.

System dependencies include:

• docker

• systemctl or init.d

• bash, curl, jq, tar

If Docker is already installed, the version is validated:

docker --version | grep "20"

If Docker is outdated, the user is prompted to upgrade before proceeding.

Step 2: Node Agent Pull

After confirming Docker, the script pulls the latest stable release of the InfraMind node agent:

docker pull inframind/node:latest

This image includes:

• The node runtime manager

• FastAPI and gRPC listeners

• Container sandboxing engine

• Telemetry and heartbeat modules

• Keypair signer and job receipt logic

The image is version-tagged and built with SHA-based reproducibility. The script verifies the digest against the published manifest and aborts if any mismatch is detected.

Step 3: Config Setup

Once the image is downloaded, the installer initializes the configuration directory:

mkdir -p ~/.inframind

The script generates a node identity keypair:

openssl genpkey -algorithm ed25519 -out ~/.inframind/identity.key
openssl pkey -in ~/.inframind/identity.key -pubout -out ~/.inframind/identity.pub

This key is used for job signing, staking, and secure communication with the scheduler mesh. The public key is registered during the first heartbeat. If a key already exists, it is reused.

The configuration file is written to ~/.inframind/config.yaml:

node_id: auto
region_hint: auto
port: 9000
container_cache: /var/lib/inframind/cache
logging: info
autoupdate: true

Advanced users may override settings by creating an inframind.env file or by using the CLI:

infra config set region_hint=us-east
infra config set logging=debug

Step 4: Background Service Start

Finally, the installer registers the node agent as a background service. If systemd is available:

sudo tee /etc/systemd/system/inframind-node.service > /dev/null <<EOF
[Unit]
Description=InfraMind Compute Node
After=network.target

[Service]
ExecStart=/usr/bin/docker run --rm --net=host \
  -v ~/.inframind:/node \
  --name inframind-node \
  inframind/node:latest
Restart=always
RestartSec=5

[Install]
WantedBy=multi-user.target
EOF

sudo systemctl daemon-reexec
sudo systemctl enable inframind-node
sudo systemctl start inframind-node

If init.d is used instead, a fallback script is created in /etc/init.d/inframind-node and registered via update-rc.d.

Once started, the node begins:

• Broadcasting heartbeats to the nearest scheduler

• Pulling model containers on demand

• Listening on port 9000 for job assignment

• Serving health and status metrics

Verify Installation

Run the following to verify status:

docker logs -f inframind-node

You should see output similar to:

[Agent] Node started with ID: 0xB71c3f...
[Agent] Listening on port 9000
[Scheduler] Connected to region mesh: europe-west
[Heartbeat] Online. Uptime: 0h0m14s

Or use the CLI:

infra status

Output:

Node ID:      0xB71c3f...
Region:       eu-west
Jobs served:  0
Cache:        1.8 GB used / 10 GB allocated
Stake:        0 INFRA

Post-Install Actions

Optionally, you may:

• Add stake: infra stake --amount 100

• Join pinning pool: infra pin --model ipfs://...

• Register for GPU jobs: infra config set allow_gpu=true

• Set location hint: infra config set region_hint=us-east

Once the node has served its first job, it will begin accruing job receipts and be visible on the public node explorer.

The entire installation process is designed to complete in under 2 minutes on most systems. Updates are handled automatically unless configured otherwise. Logs are rotated every 24 hours and stored in ~/.inframind/logs.

By the time the install script exits, the machine is a full node in the global AI compute mesh—capable of executing real inference workloads and earning rewards with no further provisioning.