Every C4 tool I've used makes you trade something: a nice canvas, or a file you can commit. Pretty UI, or git history. SaaS lock-in, or no visual editor at all.
I got tired of the tradeoff and built one that doesn't ask. Free, open source, live now 🧵
System diagrams are one of the most important foundations for the agentic software engineering.
Yet the current generation of tooling (e.g. Mermaid) feel incredibly antiquated (e.g. getting the layout right in a declarative way is close to impossible).
Are there some new tools?
Take it further with something like Structurizr and C4 — architecture as code. Now your AI isn't just generating features, it's a participant in the architectural conversation.
AI-generated architecture diagrams: pretty but possibly wrong. Must be validated against actual implementation. Evidence includes both the AI output and the human sign-off. #ProofDrivenDeliveries#CompliantDeliveries
🏴 I'll be in Edinburgh next week ... you can join me at a free session with BCS Edinburgh for a discussion of the C4 model, Structurizr "vNext", and AI.
edinburghbranch20052026.even…
If you can use TextEdit or Notepad, or if you've ever used a typewriter, you can write and publish a book on Leanpub.
It's just that easy.
Learn more in this FREE workshop this Saturday, May 16:
leanpub.com/events/book_work…
Introducing the "C4 model and Structurizr DSL pattern catalog"! 📢
This pattern catalog presents a number of minimal examples of how to use the C4 model and the Structurizr DSL to model common patterns found in software architecture.
Link: docs.structurizr.com/dsl/pat…
Prompting human code review only works if the humans have technical knowledge, which will likely decline if humans are no longer writing code.
The devs of today will become the ivory tower architects of tomorrow.
the future of software engineering seems uncontroversially prompting code review. startups will skip the code review because they’re racing against time. larger/serious orgs will take code review very seriously.
llms can do code review, but my guess is that because they have to search through large space, it will be as expensive to have say mythos review your code as it would be to have a senior dev.
based on budget:
$: prompting only
$$: low grade llm review
$$$: mid grade llm dev review
$$$$: high grade llm sr dev review
btw, software (past the bootstrapping phase) will get more expensive to make and take more time. quality will remain exactly the same as when humans were doing it: shit.
A new diagram navigator is available in the Structurizr preview build ... it provides a way to see which C4 diagrams in the workspace sit "above", "next to", and "below" the current diagram. Hope you find it useful!
@simonbrown also pointed that over reliance on ai can make developers technical skills stale.
if you stop using your own understanding, it becomes harder to validate whether the code ai generates is actually correct.
i agree with this.
the middle path is still evolving, but a simple rule works for now:
- if you use ai, read everything it generates
- if you do not understand something, learn it before using it
if you do not know what is happening then unmaintainable codebases are inevitable.
I just talked to a vibecoder on Reddit
He was actually a software developer himself, so working with AI is much easier and much better for him than a non-developer
but still...
he had issues as the codebase grew
starting from scratch works amazingly well for the first stages of the project
as things get more complicated, the prompts must be more specialised, and you have to do smaller things in smaller steps
so the gain in time you can get shrinks by a lot
I am ready to bet that the next few years will still be like the wild west, where devs use AI as companies experiment with all sorts of setups for this
we are far away from drawing conclusions but this thing will be here to stay, and we wont figure things out at least by 2030