Build Your Own, or Buy One
Most teams asking this question do not want custom software. They want their data on their own servers, or a process the tool will not bend to. Both have answers that are not a build.
It is not two options, it is four
Build versus buy is usually presented as a straight choice. In this category there are two options in the middle that solve most of what people are actually after.
Build it yourself
- Cost
- Engineering time, forever
- Control
- Total
- Time to live
- Months to years
- Upkeep
- Yours
Community Edition
- Cost
- Free, AGPL, no user limit
- Control
- Source available, modify it
- Time to live
- An afternoon to install
- Upkeep
- Yours
Self-Hosted
- Cost
- License, flat
- Control
- Your servers, your data
- Time to live
- Days
- Upkeep
- Shared, we ship updates
Cloud
- Cost
- 29, 69, or 349 dollars a month
- Control
- We host it
- Time to live
- Minutes
- Upkeep
- Ours
What building really costs
We are a software company, so we know exactly how this goes. None of these appear in the first estimate.
The build is the cheap part
A board, tasks, and a login take weeks. Permissions, notifications, audit trails, exports, search, and mobile take years, and none of them are optional once real teams depend on the tool.
It never ships once
Every operating system upgrade, framework deprecation, and security advisory becomes your ticket. Internal tools are not finished, they are staffed.
The people who built it leave
Two years in, the original authors have moved on and nobody wants to own it. This is the single most common reason internal tools get abandoned.
Opportunity cost is the real number
The engineers writing your Gantt chart are not writing the thing your customers pay you for. Price that, not just the salary.
Build if
- Your process is genuinely unlike anyone else's and is the thing you compete on.
- You need the tool embedded inside a product you already sell.
- You have a small, well defined need that will never grow, such as one form and one list.
- A license is legally impossible in your environment and no source available option exists.
Buy if
- You need boards, timelines, time, cost, and reporting, which every serious tool already has.
- You want the tool live this month rather than next year.
- You need someone else to be responsible for security patches and uptime.
- Your requirement is control over data, not control over the code. Self-hosting solves that without a build.
What you get, either way
The Build column is not a criticism. It is a list of things you would be signing up to write.
| Capability | Build | Community | Self-Hosted | Cloud |
|---|---|---|---|---|
| Live this week | ||||
| Unlimited users | You decide | |||
| Source code you can change | ||||
| Data stays on your infrastructure | ||||
| Kanban boards and tasks | You build it | |||
| Gantt, dependencies, critical path | You build it | |||
| Time tracking, budget, invoicing | You build it | |||
| Resource and workload management | You build it | Premium | ||
| Security patches maintained for you | Community | |||
| AI assistant | ||||
| Someone to call when it breaks |
Five questions before you decide
Answer these with your own numbers. We are not going to invent a build cost for you, because it depends entirely on your rates and your scope.
- 1
How many engineer months to a version people will actually use, at your fully loaded rate?
- 2
Who owns it in year three, and is that in their job description?
- 3
What is your plan for permissions, audit logging, and data export?
- 4
What happens to the tool if the two people who wrote it resign in the same quarter?
- 5
Is the requirement really custom, or is it control over where the data lives?
Build vs buy questions
Is it cheaper to build our own project management tool?+
Almost never, once you count the years after launch. The first version is the cheap part. Permissions, notifications, audit trails, exports, search, mobile, backups, and security patching are what make an internal tool expensive, and they never stop. Compare the fully loaded cost of the engineers against a flat monthly plan before you decide.
When does building actually make sense?+
When your process is the thing you compete on and is genuinely unlike anyone else's, when the tool has to be embedded inside a product you sell, or when the need is small and will never grow. If your reason is control over data rather than control over process, self-hosting gives you that without a build.
We need the data on our own servers. Do we have to build it?+
No. The Self-Hosted edition installs on your own infrastructure, private cloud, or an air gapped network, so nothing leaves your network. That is the requirement behind most build decisions in regulated and public sector organisations, and it does not need a build.
Is there a free option if we want to own the code?+
Yes. The Community Edition is free and open source under the AGPL license with no user limit. You can read the source, run it, and change it. It covers projects, tasks, and Kanban boards, and the commercial editions add the rest of the platform.
What does buying cost?+
Flat monthly plans with unlimited users: 29 dollars Basic, 69 dollars Pro, and 349 dollars Premium. Nothing is charged per seat, so the bill does not move when you hire. There is also a free forever cloud plan for one project and one user.
What if we buy and then outgrow it?+
You can move between Cloud and Self-Hosted, and the Self-Hosted edition adds modules such as document management, risk management, and release management. The REST API and MCP server are both documented, so you can build the bits that really are specific to you on top rather than rebuilding the base.
Can we buy the base and customise it?+
Yes, and that is usually the sensible middle. Custom fields, custom statuses, and per project workflows cover most of it without code. Beyond that, our professional services team builds on top of the platform, so you own the difference rather than the whole thing.
Try Before You Commit Anyone
Run the free trial for a fortnight, or download the open source edition and read the code. Both are cheaper than a discovery sprint.