The Real Cost of Running Tukuk-OS: What Is Free and What Is Not

From Tukuk-OS Wiki, the free knowledge base.

The Real Cost of Running Tukuk-OS: What Is Free and What Is Not
Type	Encyclopedia article
Subject	Tukuk-OS
Author	Fevian Donald
Published	28 October 2026
Revised	28 October 2026
License	Free to read & cite

An honest breakdown of Tukuk-OS operating costs: domain, cloud storage, open source, and hours — and why a free stack does not mean zero cost.

Contents

1. Actual Cash Expenses
2. What Comes Free
3. The Real Cost: Hours
4. Revenue Versus Cost: When Does It Make Sense?
5. What I Would Pay If It Grows
6. Lessons for Your Project

Many people are surprised to hear that a search engine can run at almost zero cost. The answer is: yes, but "almost zero" is not "zero", and there are hidden costs that are more expensive than money — time. This article breaks down what I pay, what I get for free, and where the real cost actually lives.

Actual Cash Expenses

Annual domain — this is my biggest expense. tukuk.org renews yearly at a few tens of dollars. It is also the only thing I feel is genuinely mandatory, because without a domain there is no identity.

Cloudflare (free) — DNS, HTTPS certificates, Email Routing, and Tunnel all sit on the free plan. For current traffic that plan is more than enough.

Cloudflare R2 for backups — very cheap object storage. Backup gzip files are small and written only a few times a day, so the monthly bill stays at a few cents.

Server (already owned) — Tukuk-OS runs on hardware I already have. Electricity and wear exist, but they are shared with my other computer use, so the marginal cost is barely noticeable.

Total monthly cash: almost invisible in the budget. Annually: almost entirely the domain.

What Comes Free

The software stack is almost entirely free and open source: Ubuntu, Node.js, Express, Meilisearch (community plan), systemd, and various small tools. NASA data is opened by the US government to anyone. The tools I use for hosting docs and generating the OG image are all open source.

What matters: free here means an open licence, not low quality. Meilisearch solves search problems far better than I could build in the same time, and I got it without a licence fee.

The Real Cost: Hours

This is where the real bill lives. Daily maintenance — checking logs, writing articles, adding features, testing after changes — takes far more time than any cash payment.

I estimate weekly maintenance at a few hours. If I counted that time at market rates it would easily be the largest expense of the project. But because I do the work because I want to, it is not a "cost" in the usual sense — it is an investment in the skills I gain.

Revenue Versus Cost: When Does It Make Sense?

The next question is always: does this project make enough money to cover itself? Honest answer: almost, but not quite, and that is not my main goal.

Combined, ads and other income sources cover most of the annual cash cost — in the sense that the domain and storage keep getting paid. What is not covered is the hours, but as I said, those hours are a skills investment, not an expense I expect to profit from.

I deliberately set a modest monetisation target. If I set a target of "must fully pay for itself within a year" I would be tempted to add more ads, more redirects, and more tactics that damage the reader experience. With a modest target, design decisions stay free: I add features because they are useful, not because they raise clicks.

This approach also gives me the option to say no. No sponsorships that do not fit the product's values, no misleading ads, and no privacy compromises forced by revenue needs. For a personal project the freedom to say no is probably worth more than any income anyone could offer.

What I Would Pay If It Grows

If traffic grows tenfold I would need to consider: a more powerful server (or a VPS), a wider CDN plan, and possibly a hosted search service if the index outgrows memory. I would also consider R2 for image caching and real-time notification services.

My principle: pay only when the real need arrives. Paying too early for scale you have not reached is waste; paying too late makes the site slow. The switch should happen when users — not plans — demand it.

Lessons for Your Project

If you want to build something yourself, start with this cost model: buy a domain, use open source software, host on hardware you already have, and let your costs rise with your success. That way you will not quit before starting because of a server bill, and you will not pay for something that does not need it yet.