
The fact that I can create complex diagrams just by writing code is what got me hooked in the first place. I started using Mermaid in college when I needed to make class and sequence diagrams and had no idea tools like this existed. Before this I was literally using Paint-like tools to draw boxes and arrows manually.
What I genuinely like is how fast it is to go from an idea to a proper diagram. There's no time wasted on alignment or formatting — you just write the structure and it renders. For software architecture documentation especially, this is a huge productivity boost.
The free version itself is surprisingly capable — most of what I use day to day doesn't require a paid plan at all, which is rare for tools this good.
The AI compatibility is a big plus too. Since LLMs naturally output Mermaid syntax, I can describe a system and have a diagram ready in seconds. It has become a core part of how I document architecture at work.
The integrations are what really make it stick though. GitHub and GitLab render Mermaid natively in markdown files with zero setup. Notion supports it out of the box through code blocks. VS Code renders it in the built-in markdown preview. There's even an official plugin for Confluence. The syntax is completely platform-agnostic, so the same code block works across all your tools without any changes. Review collected by and hosted on G2.com.
The syntax can have a bit of a learning curve initially, especially for non-engineers on the team. Not everyone is comfortable writing code to create a diagram, which can limit adoption beyond the engineering team.
Confluence support still requires a third-party plugin rather than being natively supported, which adds a bit of friction for teams that are heavily Confluence-based.
Also for very large and complex diagrams, the rendering can sometimes feel slow and the auto-layout doesn't always produce the cleanest output — though the visual editor does let you manually reposition nodes which helps. Review collected by and hosted on G2.com.