Install

1 Prerequisites

• Linux, x86_64 or aarch64
• wget and curl installed
• Root access — the installer creates a system user and systemd services

2 Clone the repository

git clone https://github.com/raa82/tekmerdb
cd tekmerdb

3 Run the installer

sudo ./install.sh

The installer will:

• Read the target version from Cargo.toml and download the matching pre-built binaries from the GitHub release for your architecture
• Download the ML models (~420 MB total) from HuggingFace, skipping any already present
• Create a dedicated, unprivileged tekmerdb system user
• Install everything to /opt/tekmerdb
• Install and enable three systemd units — tekmerdb-server, tekmerdb-mcp, tekmerdb-cron — enabled but not yet started

4 Start the services

sudo systemctl start tekmerdb-server tekmerdb-mcp tekmerdb-cron

The engine listens on http://127.0.0.1:3000 by default. tekmerdb-mcp and tekmerdb-cron are independent processes — restarting the engine won't take them down, and they won't restart the engine either.

5 Verify it's running

systemctl status tekmerdb-server

curl -s http://127.0.0.1:3000/status

{
  "status": "ok",
  "version": "0.x.x",
  "pfo_count": 0,
  "domain": "CriticalInfrastructure"
}

status: "ok" means the engine is accepting requests. domain should match what you set in tekmerdb-server.conf, and pfo_count tracks how many claims it holds so far.

6 Where things live