Flowcharts - Basic Syntax β
Flowcharts are composed of nodes (geometric shapes) and edges (arrows or lines). The Mermaid code defines how nodes and edges are made and accommodates different arrow types, multi-directional arrows, and any linking to and from subgraphs.
WARNING
If you are using the word "end" in a Flowchart node, capitalize the entire word or any of the letters (e.g., "End" or "END"), or apply this workaround. Typing "end" in all lowercase letters will break the Flowchart.
WARNING
If you are using the letter "o" or "x" as the first letter in a connecting Flowchart node, add a space before the letter or capitalize the letter (e.g., "dev--- ops", "dev---Ops").
Typing "A---oB" will create a circle edge.
Typing "A---xB" will create a cross edge.
A node (default) β
INFO
The id is what is displayed in the box.
TIP
Instead of flowchart
one can also use graph
.
A node with text β
It is also possible to set text in the box that differs from the id. If this is done several times, it is the last text found for the node that will be used. Also if you define edges for the node later on, you can omit text definitions. The one previously defined will be used when rendering the box.
Unicode text β
Use "
to enclose the unicode text.
Markdown formatting β
Use double quotes and backticks "` text `" to enclose the markdown text.
Direction β
This statement declares the direction of the Flowchart.
This declares the flowchart is oriented from top to bottom (TD
or TB
).
This declares the flowchart is oriented from left to right (LR
).
Possible FlowChart orientations are:
- TB - Top to bottom
- TD - Top-down/ same as top to bottom
- BT - Bottom to top
- RL - Right to left
- LR - Left to right
Node shapes β
A node with round edges β
A stadium-shaped node β
A node in a subroutine shape β
A node in a cylindrical shape β
A node in the form of a circle β
A node in an asymmetric shape β
Currently only the shape above is possible and not its mirror. This might change with future releases.
A node (rhombus) β
A hexagon node β
Parallelogram β
Parallelogram alt β
Trapezoid β
Trapezoid alt β
Double circle β
Expanded Node Shapes in Mermaid Flowcharts (v11.3.0+) β
Mermaid introduces 30 new shapes to enhance the flexibility and precision of flowchart creation. These new shapes provide more options to represent processes, decisions, events, data storage visually, and other elements within your flowcharts, improving clarity and semantic meaning.
New Syntax for Shape Definition
Mermaid now supports a general syntax for defining shape types to accommodate the growing number of shapes. This syntax allows you to assign specific shapes to nodes using a clear and flexible format:
A@{ shape: rect }
This syntax creates a node A as a rectangle. It renders in the same way as A["A"]
, or A
.
Complete List of New Shapes β
Below is a comprehensive list of the newly introduced shapes and their corresponding semantic meanings, short names, and aliases:
Semantic Name | Shape Name | Short Name | Description | Alias Supported |
---|---|---|---|---|
Process | Rectangle | rect | Standard process shape | proc , process , rectangle |
Event | Rounded Rectangle | rounded | Represents an event | event |
Terminal Point | Stadium | stadium | Terminal point | terminal , pill |
Subprocess | Framed Rectangle | fr-rect | Subprocess | subprocess ,subproc , framed-rectangle , subroutine |
Database | Cylinder | cyl | Database storage | db , database , cylinder |
Start | Circle | circle | Starting point | circ |
Odd | Odd | odd | Odd shape | |
Decision | Diamond | diam | Decision-making step | decision , diamond |
Prepare Conditional | Hexagon | hex | Preparation or condition step | hexagon , prepare |
Data Input/Output | Lean Right | lean-r | Represents input or output | lean-right , in-out |
Data Input/Output | Lean Left | lean-l | Represents output or input | lean-left , out-in |
Priority Action | Trapezoid Base Bottom | trap-b | Priority action | priority , trapezoid-bottom |
Manual Operation | Trapezoid Base Top | trap-t | Represents a manual task | manual , trapezoid-top |
Stop | Double Circle | dbl-circ | Represents a stop point | double-circle |
Text Block | Text Block | text | Text block | - |
Card | Notched Rectangle | notch-rect | Represents a card | card , notched-rectangle |
Lined/Shaded Process | Lined Rectangle | lin-rect | Lined process shape | lined-rectangle ,lined-process , lin-proc ,shaded-process |
Start | Small Circle | sm-circ | Small starting point | start , small-circle |
Stop | Framed Circle | fr-circ | Stop point | stop , framed-circle |
Fork/Join | Filled Rectangle | fork | Fork or join in process flow | join |
Collate | Hourglass | hourglass | Represents a collate operation | hourglass |
Comment | Curly Brace | brace | Adds a comment | comment , brace-l |
Comment Right | Curly Brace | brace-r | Adds a comment | - |
Comment with braces on both sides | Curly Braces | braces | Adds a comment | - |
Com Link | Lightning Bolt | bolt | Communication link | com-link , lightning-bolt |
Document | Document | doc | Represents a document | doc , document |
Delay | Half-Rounded Rectangle | delay | Represents a delay | half-rounded-rectangle |
Direct Access Storage | Horizontal Cylinder | h-cyl | Direct access storage | das , horizontal-cylinder |
Disk Storage | Lined Cylinder | lin-cyl | Disk storage | disk , lined-cylinder |
Display | Curved Trapezoid | curv-trap | Represents a display | curved-trapezoid , display |
Divided Process | Divided Rectangle | div-rect | Divided process shape | div-proc , divided-rectangle , divided-process |
Extract | Triangle | tri | Extraction process | extract , triangle |
Internal Storage | Window Pane | win-pane | Internal storage | internal-storage , window-pane |
Junction | Filled Circle | f-circ | Junction point | junction , filled-circle |
Lined Document | Lined Document | lin-doc | Lined document | lined-document |
Loop Limit | Trapezoidal Pentagon | notch-pent | Loop limit step | loop-limit , notched-pentagon |
Manual File | Flipped Triangle | flip-tri | Manual file operation | manual-file , flipped-triangle |
Manual Input | Sloped Rectangle | sl-rect | Manual input step | manual-input , sloped-rectangle |
Multi-Document | Stacked Document | docs | Multiple documents | documents , st-doc , stacked-document |
Multi-Process | Stacked Rectangle | st-rect | Multiple processes | procs , processes , stacked-rectangle |
Paper Tape | Flag | flag | Paper tape | paper-tape |
Stored Data | Bow Tie Rectangle | bow-rect | Stored data | stored-data , bow-tie-rectangle |
Summary | Crossed Circle | cross-circ | Summary | summary , crossed-circle |
Tagged Document | Tagged Document | tag-doc | Tagged document | tag-doc , tagged-document |
Tagged Process | Tagged Rectangle | tag-rect | Tagged process | tagged-rectangle ,tag-proc , tagged-process |
Example Flowchart with New Shapes β
Hereβs an example flowchart that utilizes some of the newly introduced shapes:
Process β
Event β
Terminal Point (Stadium) β
Subprocess β
Database (Cylinder) β
Start (Circle) β
Odd β
Decision (Diamond) β
Prepare Conditional (Hexagon) β
Data Input/Output (Lean Right) β
Data Input/Output (Lean Left) β
Priority Action (Trapezoid Base Bottom) β
Manual Operation (Trapezoid Base Top) β
Stop (Double Circle) β
Text Block β
Card (Notched Rectangle) β
Lined/Shaded Process β
Start (Small Circle) β
Stop (Framed Circle) β
Fork/Join (Long Rectangle) β
Collate (Hourglass) β
Comment (Curly Brace) β
Comment Right (Curly Brace Right) β
Comment with braces on both sides β
Com Link (Lightning Bolt) β
Document β
Delay (Half-Rounded Rectangle) β
Direct Access Storage (Horizontal Cylinder) β
Disk Storage (Lined Cylinder) β
Display (Curved Trapezoid) β
Divided Process (Divided Rectangle) β
Extract (Small Triangle) β
Internal Storage (Window Pane) β
Junction (Filled Circle) β
Lined Document β
Loop Limit (Notched Pentagon) β
Manual File (Flipped Triangle) β
Manual Input (Sloped Rectangle) β
Multi-Document (Stacked Document) β
Multi-Process (Stacked Rectangle) β
Paper Tape (Flag) β
Stored Data (Bow Tie Rectangle) β
Summary (Crossed Circle) β
Tagged Document β
Tagged Process (Tagged Rectangle) β
Links between nodes β
Nodes can be connected with links/edges. It is possible to have different types of links or attach a text string to a link.
A link with arrow head β
An open link β
Text on links β
or
A link with arrow head and text β
or
Dotted link β
Dotted link with text β
Thick link β
Thick link with text β
An invisible link β
This can be a useful tool in some instances where you want to alter the default positioning of a node.
Chaining of links β
It is possible declare many links in the same line as per below:
It is also possible to declare multiple nodes links in the same line as per below:
You can then describe dependencies in a very expressive way. Like the one-liner below:
If you describe the same diagram using the basic syntax, it will take four lines. A word of warning, one could go overboard with this making the flowchart harder to read in markdown form. The Swedish word lagom
comes to mind. It means, not too much and not too little. This goes for expressive syntaxes as well.
New arrow types β
There are new types of arrows supported:
- circle edge
- cross edge
Circle edge example β
Cross edge example β
Multi directional arrows β
There is the possibility to use multidirectional arrows.
Minimum length of a link β
Each node in the flowchart is ultimately assigned to a rank in the rendered graph, i.e. to a vertical or horizontal level (depending on the flowchart orientation), based on the nodes to which it is linked. By default, links can span any number of ranks, but you can ask for any link to be longer than the others by adding extra dashes in the link definition.
In the following example, two extra dashes are added in the link from node B to node E, so that it spans two more ranks than regular links:
Note Links may still be made longer than the requested number of ranks by the rendering engine to accommodate other requests.
When the link label is written in the middle of the link, the extra dashes must be added on the right side of the link. The following example is equivalent to the previous one:
For dotted or thick links, the characters to add are equals signs or dots, as summed up in the following table:
Length | 1 | 2 | 3 |
---|---|---|---|
Normal | --- | ---- | ----- |
Normal with arrow | --> | ---> | ----> |
Thick | === | ==== | ===== |
Thick with arrow | ==> | ===> | ====> |
Dotted | -.- | -..- | -...- |
Dotted with arrow | -.-> | -..-> | -...-> |
Special characters that break syntax β
It is possible to put text within quotes in order to render more troublesome characters. As in the example below:
Entity codes to escape characters β
It is possible to escape characters using the syntax exemplified here.
Numbers given are base 10, so #
can be encoded as #35;
. It is also supported to use HTML character names.
Subgraphs β
subgraph title
graph definition
end
An example below:
You can also set an explicit id for the subgraph.
flowcharts β
With the graphtype flowchart it is also possible to set edges to and from subgraphs as in the flowchart below.
Direction in subgraphs β
With the graphtype flowcharts you can use the direction statement to set the direction which the subgraph will render like in this example.
Limitation β
If any of a subgraph's nodes are linked to the outside, subgraph direction will be ignored. Instead the subgraph will inherit the direction of the parent graph:
Markdown Strings β
The "Markdown Strings" feature enhances flowcharts and mind maps by offering a more versatile string type, which supports text formatting options such as bold and italics, and automatically wraps text within labels.
Formatting:
- For bold text, use double asterisks (
**
) before and after the text. - For italics, use single asterisks (
*
) before and after the text. - With traditional strings, you needed to add
<br>
tags for text to wrap in nodes. However, markdown strings automatically wrap text when it becomes too long and allows you to start a new line by simply using a newline character instead of a<br>
tag.
This feature is applicable to node labels, edge labels, and subgraph labels.
The auto wrapping can be disabled by using
---
config:
markdownAutoWrap: false
---
graph LR
Interaction β
It is possible to bind a click event to a node, the click can lead to either a javascript callback or to a link which will be opened in a new browser tab.
INFO
This functionality is disabled when using securityLevel='strict'
and enabled when using securityLevel='loose'
.
click nodeId callback
click nodeId call callback()
- nodeId is the id of the node
- callback is the name of a javascript function defined on the page displaying the graph, the function will be called with the nodeId as parameter.
Examples of tooltip usage below:
<script>
window.callback = function () {
alert('A callback was triggered');
};
</script>
The tooltip text is surrounded in double quotes. The styles of the tooltip are set by the class .mermaidTooltip
.
Success The tooltip functionality and the ability to link to urls are available from version 0.5.2.
?> Due to limitations with how Docsify handles JavaScript callback functions, an alternate working demo for the above code can be viewed at this jsfiddle.
Links are opened in the same browser tab/window by default. It is possible to change this by adding a link target to the click definition (_self
, _blank
, _parent
and _top
are supported):
Beginner's tipβa full example using interactive links in a html context:
<body>
<pre class="mermaid">
flowchart LR
A-->B
B-->C
C-->D
click A callback "Tooltip"
click B "https://www.github.com" "This is a link"
click C call callback() "Tooltip"
click D href "https://www.github.com" "This is a link"
</pre>
<script>
window.callback = function () {
alert('A callback was triggered');
};
const config = {
startOnLoad: true,
flowchart: { useMaxWidth: true, htmlLabels: true, curve: 'cardinal' },
securityLevel: 'loose',
};
mermaid.initialize(config);
</script>
</body>
Comments β
Comments can be entered within a flow diagram, which will be ignored by the parser. Comments need to be on their own line, and must be prefaced with %%
(double percent signs). Any text after the start of the comment to the next newline will be treated as a comment, including any flow syntax
Styling and classes β
Styling links β
It is possible to style links. For instance, you might want to style a link that is going backwards in the flow. As links have no ids in the same way as nodes, some other way of deciding what style the links should be attached to is required. Instead of ids, the order number of when the link was defined in the graph is used, or use default to apply to all links. In the example below the style defined in the linkStyle statement will belong to the fourth link in the graph:
linkStyle 3 stroke:#ff3,stroke-width:4px,color:red;
It is also possible to add style to multiple links in a single statement, by separating link numbers with commas:
linkStyle 1,2,7 color:blue;
Styling line curves β
It is possible to style the type of curve used for lines between items, if the default method does not meet your needs. Available curve styles include basis
, bumpX
, bumpY
, cardinal
, catmullRom
, linear
, monotoneX
, monotoneY
, natural
, step
, stepAfter
, and stepBefore
.
In this example, a left-to-right graph uses the stepBefore
curve style:
%%{ init: { 'flowchart': { 'curve': 'stepBefore' } } }%%
graph LR
For a full list of available curves, including an explanation of custom curves, refer to the Shapes documentation in the d3-shape project.
Styling a node β
It is possible to apply specific styles such as a thicker border or a different background color to a node.
Classes β
More convenient than defining the style every time is to define a class of styles and attach this class to the nodes that should have a different look.
A class definition looks like the example below:
classDef className fill:#f9f,stroke:#333,stroke-width:4px;
Also, it is possible to define style to multiple classes in one statement:
classDef firstClassName,secondClassName font-size:12pt;
Attachment of a class to a node is done as per below:
class nodeId1 className;
It is also possible to attach a class to a list of nodes in one statement:
class nodeId1,nodeId2 className;
A shorter form of adding a class is to attach the classname to the node using the :::
operator as per below:
This form can be used when declaring multiple links between nodes:
CSS classes β
It is also possible to predefine classes in CSS styles that can be applied from the graph definition as in the example below:
Example style
<style>
.cssClass > rect {
fill: #ff0000;
stroke: #ffff00;
stroke-width: 4px;
}
</style>
Example definition
Default class β
If a class is named default it will be assigned to all classes without specific class definitions.
classDef default fill:#f9f,stroke:#333,stroke-width:4px;
Basic support for fontawesome β
It is possible to add icons from fontawesome.
The icons are accessed via the syntax fa:#icon class name#.
Mermaid supports Font Awesome if the CSS is included on the website. Mermaid does not have any restriction on the version of Font Awesome that can be used.
Please refer the Official Font Awesome Documentation on how to include it in your website.
Adding this snippet in the <head>
would add support for Font Awesome v6.5.1
<link
href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.1/css/all.min.css"
rel="stylesheet"
/>
Custom icons β
It is possible to use custom icons served from Font Awesome as long as the website imports the corresponding kit.
Note that this is currently a paid feature from Font Awesome.
For custom icons, you need to use the fak
prefix.
Example
flowchart TD
B[fa:fa-twitter] %% standard icon
B-->E(fak:fa-custom-icon-name) %% custom icon
And trying to render it
Graph declarations with spaces between vertices and link and without semicolon β
In graph declarations, the statements also can now end without a semicolon. After release 0.2.16, ending a graph statement with semicolon is just optional. So the below graph declaration is also valid along with the old declarations of the graph.
A single space is allowed between vertices and the link. However there should not be any space between a vertex and its text and a link and its text. The old syntax of graph declaration will also work and hence this new feature is optional and is introduced to improve readability.
Below is the new declaration of the graph edges which is also valid along with the old declaration of the graph edges.
Configuration β
Renderer β
The layout of the diagram is done with the renderer. The default renderer is dagre.
Starting with Mermaid version 9.4, you can use an alternate renderer named elk. The elk renderer is better for larger and/or more complex diagrams.
The elk renderer is an experimental feature. You can change the renderer to elk by adding this directive:
%%{init: {"flowchart": {"defaultRenderer": "elk"}} }%%
INFO
Note that the site needs to use mermaid version 9.4+ for this to work and have this featured enabled in the lazy-loading configuration.
Width β
It is possible to adjust the width of the rendered flowchart.
This is done by defining mermaid.flowchartConfig or by the CLI to use a JSON file with the configuration. How to use the CLI is described in the mermaidCLI page. mermaid.flowchartConfig can be set to a JSON string with config parameters or the corresponding object.
mermaid.flowchartConfig = {
width: 100%
}