Software that does one job well and stays yours to run.
Not every problem is a website problem. Sometimes the thing you need is a tool: something that tracks inventory across sites, keeps records straight, runs a process nobody wants to do by hand. I build those as their own software rather than bolting them onto a page.
Expected outcome
- A process that used to live in one person's head runs on its own.
- The tool keeps working without a subscription to somebody else's platform.
- You can read what it does, and so can whoever maintains it next.
- It does one job, so it can be finished rather than endlessly extended.
01
What this covers
Software that stands on its own, as opposed to the admin screens attached to a website. Both get built; they are scoped differently.
- Internal tools for work that currently happens in a spreadsheet
- Local-first apps that keep working when the connection does not
- Operations and records systems across more than one location
- Research and reporting tools that produce something you can act on
- Small utilities that do a single job and then get out of the way
02
How it gets built
Small software survives on being understandable a year later, by someone who was not there when it was written. That decides most of the technical choices.
- Boring where boring works, because cleverness has a carrying cost
- State you can see, so the thing can be debugged by looking at it
- Automation you can interrupt, rather than automation you have to trust
- Written down as it is built, so picking it back up is not an act of memory
03
Scope and investment
Tool work is scoped from the process it replaces, not from a feature list, because the honest size only shows up once the current workaround is visible.
- A single focused tool usually starts around $2,500
- Multi-location or records-heavy systems run $10,000–$25,000+
- Scoped from a walkthrough of how the work happens today
- You keep the code, the data, and the ability to hand it to someone else
Where this has been built
The opinions above came from somewhere.
None of these is a product you can buy. They are the working evidence behind the section above — software built for my own operations first, which is where the views on how it should be built come from.
Day-to-day operations
Business operations
Planning, sales follow-up, records, events, rentals, and the operating queues behind a venue or a logistics run.
Nine tools. Three of them run here day to day; the rest are in development.
Research and review
Research and decision support
Market research, structured reasoning, risk review, and using AI in a way that can be audited afterwards rather than taken on faith.
Six tools, two of them still experimental.
Finance and records
Records and bookkeeping
Bookkeeping, billing, customer records, and keeping location information straight when the same business exists in several systems at once.
Three tools, in development.
Recording and live operations
Capture and streaming
Screen and audio capture, live stream operations, chat handling, and the service that keeps those pieces talking to each other.
Four desktop apps and one worker service, in development.
Practical local tools
Small utilities
Local coordination, private matching, regional time handling, and storage planning. Single-purpose things that do one job and stop.
Four tools, two of them still experimental.
Have a process that has outgrown a spreadsheet?
Describe what it does now, including the parts that only work because you remember them. That is usually the shape of the tool.