Expressions

Expressions are a way to reference data in your workflows. To use an expression, you must wrap an expression context around:

${{ <expression> }}

Expressions are evaluated into values before the start of each action run. Expressions are primarily used within action inputs to reference and manipulate data.

​

Expression Contexts

Tracecat comes with five expression context types:

​

Actions

Use the ACTIONS prefix to reference the result of an action.

{"message": "hello world"}

${{ ACTIONS.welcome_user.result.data.message }}

​

Trigger

Workflows can be triggered via webhook, manual trigger, or the Execute Child Workflow action. Use the TRIGGER prefix to reference the data from the trigger as a JSON object.

${{ TRIGGER.<jsonpath> }}

export WEBHOOK_URL=https://tracecat.com/api/webhooks/<webhook_id>/<secret>
curl -X POST $WEBHOOK_URL \
    -H "Content-Type: application/json" \
    -d '{"data": {"name": "John", "age": 30}}'

${{ TRIGGER.data.name }}

​

Secrets

Secrets stored in the secrets manager can be accessed using the SECRETS prefix:

${{ SECRETS.<name>.<key> }}

​

Functions

Use functions to manipulate data. Functions are called using the FN prefix. For example, the following expression will return the boolean value true:

${{ FN.contains("a", ["a", "b", "c"]) }}

We also support nested function and expression calls.

For example, the following expression will:

1. First evaluate the expression FN.join(["hello", "world"], " ") as hello world
2. Then call FN.contains with the arguments hello world and ["hello", "world", "tracecat"]
3. And finally, return the boolean value true

${{ FN.contains(FN.join(["hello", "world"], " "), ["hello", "world", "tracecat"]) }}

​

Variables

Variables are accessed using the VARS prefix:

${{ VARS.<jsonpath> }}

${{ VARS.api_url }}

​

Operators

We support standard operators (+, -, *, /) for all Pythonobject types (e.g. int, float, str, datetime, timedelta, list, dict).

${{ 1 + 2 }}

${{ "hello " + "world" }}

​

Typecasting

We also support casting expressions into different data types with the following syntax:

Supported typecasts: