"The Changelog as Proof That Buy-Once Does Not Mean Abandoned"

#The trust problem with buy once

There is a reasonable fear that comes with buying software once, and I want to name it directly, because if I do not, it will sit in the back of your mind while you read everything else on this site.

The fear is: what if they stop?

Subscription software has a built-in answer to that fear. As long as you are paying, they have a reason to keep working on it. The incentive is aligned. Buy-once software does not have that. You paid once. The incentive to keep shipping has to come from somewhere else, and if that somewhere else runs out, you own a tool that slowly goes stale.

It is a real fear. I have lived it as a user, buying tools that were great on launch day and untouched two years later. So I am not going to tell you to ignore it. I am going to show you the changelog and let you judge for yourself whether this product is alive.

#What the changelog actually says

I am going to walk through the last several owntime releases, not as marketing, but as evidence. These are real releases, on real dates, with real notes. Read them and decide for yourself what they say about whether this is a product someone is actively maintaining.

**v0.4.3, August 30

.** A billing math fix. Invoices built from timer entries could show sixty times the real amount, because a rounding function returned minutes where hours were expected. That is the kind of bug that makes a product unusable for its core job, and it was found and fixed. The same release fixed a CSV import bug where "90m" parsed as ninety hours instead of one and a half. It also added a test gate, eighteen tests covering [invoice](/) math, rounding, rate history, CSV round-trips, and trial days, so those specific bugs do not come back.

v0.4.2, August 28. A visual polish pass. Buttons using the wrong font, shortcut badges not styled, dark-mode chips broken. No behavior changes. The kind of release that says "I care about how this looks, not just whether it works."

v0.4.0, August 25. Timeline, the opt-in passive tracker that records your frontmost window locally. Own-bucket backup to your R2 or S3 on every save. Budgets going live with header bars at 75, 90, and 100 percent. A pomodoro and reminder system with quiet hours. This was not a small release.

v0.3.0, August 24. Window state memory, donut charts in reports, a CLI that talks to the app on localhost for the Raycast crowd. Drag entries onto projects. Drop a CSV anywhere to import. Dark mode.

v0.2.0 and earlier. The foundation. The weekly grid, clients and projects and tasks, the invoice builder, the timer with a global hotkey, the thing that became the product.

Read that sequence again. Five releases in about a week, each one real, each one fixing or building something specific. That is not the cadence of a product someone built and abandoned. That is the cadence of a product someone is actively, even obsessively, working on.

#What the changelog does not say

I want to be fair about what reading a changelog can and cannot tell you.

It can tell you the product is alive right now. It cannot tell you it will be alive in two years. Nothing can tell you that, not for a subscription tool and not for a buy-once one. Harvest was actively maintained for twenty years and then got acquired and repriced. Active maintenance is not a guarantee. It is just evidence.

It can tell you the developer fixes bugs quickly. The sixty-times-billing-math bug in v0.4.3 is a good example. It cannot tell you there are no bugs you have not found yet. There are always bugs you have not found yet.

It can tell you the pace of shipping. It cannot tell you the pace will continue. Life happens. Solo developers get sick, get busy, get bored. I am shipping fast right now. I cannot promise I will ship fast forever.

What I can tell you is that even if the shipping stops, the tool does not stop working. That is the buy-once guarantee, and it is the reason I can be honest about the risk of abandonment without it being a dealbreaker. The worst case is the changelog goes quiet and your copy keeps working. The best case is it does not.

#How to use this

If you are evaluating owntime, do not take my word that it is maintained. Look at the changelog yourself. Check the dates. Check whether the releases are real fixes and features or vague "performance improvements." Check whether the pace is one you trust.

And then check the thing that actually matters for your decision: does the current version do what you need? Because if it does, the question of whether I will still be shipping in 2028 is less important than you think. You are buying the tool as it exists today, with the guarantee that it will keep working tomorrow whether I ship again or not.

The changelog is proof of life. The buy-once license is proof of survival. You need both, and owntime has both.

Read the full changelog · Try owntime free for 14 days · See pricing, one time

---