How to Use Mermaid in Microsoft Loop Easily 2026

Click onto an empty line in your Loop page and type the slash command for Mermaid, which swaps that line for a code entry module instead of a plain text block. This is the same slash-command pattern Loop uses for tables, images, and other structured content, so it feels familiar once you've used it once. The block opens with a language selector already set, visible in the corner of the module, confirming Mermaid is the active language, not a generic code block.

A minimal flowchart starts with a direction declaration like flowchart TD for top-down, followed by node definitions and arrows on the next lines. Something as short as A[Need Diagrams?] -->|use| B[Mermaid] defines two labeled boxes and an arrow between them tagged with the word "use": three lines of plain text producing a fully rendered flowchart with no drawing involved.

Node names on the left of the brackets (A, B) are just internal identifiers; the text inside the brackets is what shows up in the box, so labels can be renamed freely without breaking the connections between them. Those identifiers are case-sensitive, and a typo in one creates a new disconnected node instead of throwing an error, which is worth knowing before you spend time hunting for a missing arrow.

Loop renders the Mermaid block as a code editor on top with line numbers down the left side, and the compiled diagram appears directly underneath it in the same container, with small toolbar icons and a language dropdown along the top edge. There's no separate preview pane to hunt for; editing the code and watching the shape change happen in the same visual space, which makes it easy to iterate on a diagram by trial and error. Scrolling the page keeps the code and its rendered output together as one unit instead of splitting them apart.

Once the basic two-box flowchart works, add more lines following the same pattern: another node definition and arrow per line, to build out a longer process. Mermaid also supports other diagram types beyond flowcharts, including sequence diagrams for showing interactions over time and Gantt charts for timelines, all written with their own compact syntax. Because everything is plain text, restructuring a diagram is just editing lines instead of dragging shapes around a canvas, which is faster once you're comfortable with the syntax.

Because the diagram lives as a code block on a shared Loop page instead of an exported image, any teammate with edit access can open it and change a label, add a branch, or fix a typo directly in the syntax. The rendered diagram updates immediately for whoever is looking at the page, without anyone needing to re-export or re-upload anything. The source and the visual are the same object, so they can never drift out of sync the way an external drawing file and its exported copy can.