Table of contents

Simple signals

Draw a signal from one participant to another like this. The participants are automatically created when they are used.

To include a colon in a participant name, enclose it in quotes.

":Alice"->":Bob": Hello
        

# This is a comment.

Alice->Bob: Filled arrow
Alice->>Bob: Open arrow
Bob-->Alice: Dotted line
Bob-->>Alice: Dotted Line, open arrow
Bob-->Alice: Double arrow
Alice->(1)Bob: Slightly delayed
Bob->(5)Alice: Highy delayed

Changing the order of participants

If you want to participants to be shown in a different order than they are used, declare them first using the participant keyword. You can also rename them this way to save typing.

participant Bob
participant Alice
participant "I have a really\nlong name" as L

Alice->Bob: Authentication Request
Bob->Alice: Authentication Response
Bob->L: Log transaction

Actors

Declare a participant using the actor keyword and it will appear with a person drawn above it.

actor User
participant Terminal

User->Terminal: Login

Signal to self

A participant can send a signal to itself. This will result in an arrow that turns back on itself.

You may break the text into multiple lines by using "\n".

Alice->Alice: This is a signal to self.\nIt also demonstrates \nmultiline \ntext.

Grouping signals together

You can group signals together using the alt/else, opt, loop, par, and seq keywords. All of them can take a text description that will be displayed in the group header. Use the end keyword to signal the end of a group. The groups may be nested to any depth.

The else part is optional and can be repeated more than once for each condition.

Alice->Bob: Authentication Request
alt successful case
Bob->Alice: Authentication Accepted
else some kind of failure
Bob->Alice: Authentication Failure
opt
    loop 1000 times
        Alice->Bob: DNS Attack
    end
end
else Another type of failure
Bob->Alice: Please repeat
end

Notes in the diagram

You can add notes to the diagram. Notes can be placed to the left or right of a participant, or placed over one or more participants.

If a note contains more than one line, it will not be word-wrapped. Instead, it will be formatted exacty as written.

participant Alice
participant Bob

note left of Alice 
This is displayed 
left of Alice.
end note
note right of Alice: This is displayed right of Alice.
note over Alice: This is displayed over Alice.
note over Alice, Bob: This is displayed over Bob and Alice.

Extended text descriptions (premium)

Subscribers can explain sequences that follow simply by indenting your explanations with a space.

 You can explain the sequences that follow
 simply by indenting your explanations 
 with a space.
       
Alice->Bob: Wow!

Lifeline activation and destruction

Use the + and - with signals to denote object activation. While activated, the participant's lifeline will be highlighted. + will activate the receiver, and - will deactivate the sender.

Use * with signals to create a participant.

A->*B: hello.

You can use the destroy keyword to destroy a participant. The participant's lifeline will end at the previous signal.

User->+A: DoWork
A->*+B: <<createRequest>>
B->*+C: DoWork
C-->B: WorkDone
destroy C
B-->-A: RequestCreated
A->User: Done

Alternate syntax

Ge more control over activations using the activate and deactive keywords. The commands apply to the previous signal.

Alice->Bob: Do some work!
activate Bob

Bob->Bob: Work routine
activate Bob
deactivate Bob

Bob->Alice: All done!
deactivate Bob

Only arrows can cause activations

Activations and deactivations attach themselves to the previoius signal. Neigher a note nor a state can cause a deactivation. If you want to deactivate spontaneously, use a signal-to-self.

A->+B: Activate please
B->-B: I'm deactivating now\n by myself

One arrow can activate/deactivate many participants at a time.

Using the explict syntax gives more flexibility.

A->B: 
activate A
activate B

B->C: 
deactivate A
deactivate B
activate C

C->C:
deactivate C