A console for the team that runs your servers
sysx is a self-hosted operations console for small teams looking after a fleet of Linux servers — inventory, recurring maintenance, health monitoring, and an AI assistant that reads the telemetry so you don't have to, all in one place, on hardware you control.
Built for the people actually doing the work
sysx began in-house, solving a single problem: how to keep sysadmins in different timezones working from the same picture. What it became was decided by the people using it. Servers needed maintenance checks — some far more often than others. Hours needed tracking against clients for billing. Every request was weighed against whether it earned its place, so what grew was a coherent server, task and billing system rather than a pile of features. It has been in production on our own estate ever since. We have now made it available for other teams to deploy.
The restraint is deliberate. There is no time-series database, no dashboard-of-dashboards, no per-host licence. sysx covers the part those platforms leave you to build yourself: the maintenance schedule, the institutional knowledge, and the plain-English record of what happened and why.
What it does
Inventory and accounts
Every server, grouped by client, with role notes, OS, IP, and a plain-English description of what "normal" looks like on that machine. Projects and client records sit alongside the server list. Monthly hour budgets per server for billing and capacity planning.
Recurring maintenance
Reusable task types — SSL checks, backup verification, log rotation, failover tests, lockdown reviews — each with a frequency per server. Tasks generate on schedule and get allocated across the team, capped per person. Every task gets a report: findings, an outcome, time spent. Comments thread on servers, tasks and clients.
Health monitoring
Two ways to collect data, so every box is covered: a shell agent that pushes from cron, or a PHP endpoint polled over HTTPS. sysx tracks CPU load, memory, per-mount disk usage and listening ports, and builds a rolling baseline of what normally listens — so it notices when a service stops. Threshold checks are deterministic, written in plain code, and are the authoritative layer.
AI assistance
A scheduled review reads the collected data and posts a short summary to your team chat a few times a day — and stays quiet when there's nothing to report. A chat assistant answers questions about any server, reading live telemetry, the maintenance record, and your team's own notes. A written knowledge base holds what telemetry can't show, and every entry is reviewable and correctable by an admin.
The full breakdown is on the features page.
Why it's built this way
The checks are deterministic, and senior to the AI
Threshold breaches are decided in code, not by a model. The AI describes and prioritises — it is never the only thing standing between a full disk and a silent all-clear.
It knows what a server is for
Not just the numbers — the context. A scheduled job that fills a third of a disk overnight gets described as expected, not reported as an emergency.
Institutional memory stays put
Maintenance reports and working notes are captured and stay reachable, so knowledge doesn't leave when someone does.
It's small enough to read
Plain PHP 8 and MySQL, no framework, no build pipeline. You can read the whole thing, and it will still run in five years.
Runs on what you've already got
sysx is plain PHP and MySQL. No framework, no build step, no container orchestration. It runs comfortably on a modest VPS alongside your other sites.
Interested in running sysx on your own estate?
Tell us how many servers you look after and who looks after them, and we'll take it from there.
Enquire about a deployment