Architecture docs that live in a wiki die in a wiki. The docs that get read are the ones in the repo — and the docs that get written are the ones the scaffold already included.
Most architecture docs are written after the fact, in a separate tool, by someone who wasn't there. They rot. C4 — Context, Container, Component, Code — works because it's structured, and it works in a repo because it lives next to the code it describes.
First Things: Docs Before Code
Begin with the end in mind: what should a new dev understand in ten minutes? The C4 model answers that — the context the system operates in, the containers it's made of, the components inside them, and the code-level invariants. If those are in the repo, onboarding is a README away.
- Context: what the system talks to and why.
- Container: the deployable units and their boundaries.
- Component: the pieces inside each container.
- Code: the invariants — the decisions that don't change.
Issue-Tree the Rot
Why do architecture docs rot? They live apart from the code, they're written once, and nobody updates them. C4 in the repo fixes all three: it's next to the code, it's structured, and the invariants section is the part that should change least.
What the Templates Ship
Every FACTA Template ships its own README, architecture docs (C4 + invariants), and skeleton. The docs aren't a phase two — they're baked in, so the agent and the human both inherit the architecture instead of guessing it.
Who It's For
This is for the small team that can't afford a dedicated architect but can't afford architectural drift either. C4 in the repo is the cheapest architectural discipline that actually survives — because it lives where the code lives.
Conclusion
Architecture docs work when they're structured, in the repo, and inherited from the scaffold. C4 + invariants baked into every template is the cheapest way to give a small team the discipline of a much larger one.
About FACTA
FACTA helps startups and growth-stage teams turn AI into production systems that keep running — not demos that impress once.
We design the architecture around the parts that actually break under real usage: tooling you own, credentials you control, failover, cost controls, observability. The boring infrastructure that keeps a system alive after launch.
Led by Matías Baglieri and Carolina Fogliato, we focus on one thing:
AI leadership that builds. Not just advises.
Copy a template and get the C4 docs with it.
See the invariants post for why the invariants section matters most.
Explore AI Automation
