Starhold Software · automation and platform engineering
Automation and platform engineering for teams that outgrew scripts.
You wrote a script. Then a folder of them. Now something that matters depends on them — I build the version that holds, and I keep it running.
Things you can check before you read another word
What I sell
Four ways to hire me
Scoped and quoted per job. No retainer, no minimum, no discovery-phase invoice.
Automation
The manual step, removed. Python, PowerShell, and Bash that runs on a schedule, records what it did, and says so when it fails.
Platform and HA hosting
Somewhere for it to live. Rootless containers, private mesh networking, backups, and a status page anyone can open.
Custom bots
Discord run like business software, under your server's own name and avatar: moderation, roles, logging, and automation. An Ops Pack adds server, application, and scheduled-scraping monitoring.
Full stack solutions
A whole application, not a script. Proven open-source components, assembled and operated as one system scoped to your business — the job a multi-thousand-dollar platform does, without that invoice.
The fleet
Selected work
Four things I built and still run. Each one is here because of what it proves, not because it is for sale.
Shushgame
A Discord server that is itself the game, with its own site and its own payments.
Live
A whole product carried end to end: design, bot, billing, hosting. Nothing about it was handed to someone else.
Custom Discord Bots
A configurable bot sold as a flat monthly subscription, under your server name.
Live
Live Stripe Checkout, webhook signatures verified before anything acts on them, and a catalog published as machine-readable JSON that my own Discord bot consumes as a contract.
QNix
Bring-your-own-device hosting with high availability.
In development
Architecture, hosting, and tenancy and security notes were written down and published before the product existed.
Starhold itself
This site, the docs, the status page, and the machines underneath them.
Live
A dedicated server and a VPS on a private Tailscale mesh, running rootless Podman containers, with 112 unit tests and 25 browser tests in this repository.
How I work
Operating principles
Short enough to remember, specific enough to hold me to.
- Uptime is published, not claimed. Every service I run reports to a status page you can open right now, including when it is down.
- Nothing is rented that can be owned. The hosts are mine to administer, joined by a private mesh, and the services run rootless.
- Tests ship with the work. This storefront alone carries 112 unit tests and 25 browser tests, and the counts on this page are generated from the repository rather than typed in. Read them →
- Money is verified, not trusted. Payments go through Stripe Checkout, and every webhook signature is checked before a single order is acted on.
- Design notes come before code. The architecture for the platform in development is written up and published where you can read it.
Who you are hiring
Kyle Green
I run Starhold Software. It is one person, and that person is me.
- AWS Certified Solutions Architect
- Certified in Python
- Certified in log aggregation at the administration tier
- By day I keep automation running for an organisation that operates worldwide. I do not name the employer here, and this work is entirely separate from it.
Next step
Tell me what you keep doing by hand.
Send me the process, the two systems that refuse to talk to each other, or the thing that keeps breaking overnight. I will tell you what it takes to fix it and what it costs.
Currently taking work for Q4 2026.