Variables

Extracting outputs, template interpolation, built-in variables, and the Faker data library.

TestMesh has a single variable scope shared across all steps in a flow. Variables come from four sources: the env block, system built-ins, Faker data generators, and step outputs.

Template syntax

Variables are referenced with ${VARIABLE_NAME} in any string field inside a step's config, assert, or output blocks.

config:
  url: "${API_URL}/users/${user_id}"
  body:
    email: "${USER_EMAIL}"

1. env variables

Defined in the flow's env block. Values are strings that can themselves reference system environment variables using the same ${...} syntax.

flow:
  env:
    API_URL: "${API_BASE_URL}/v1"    # References a shell environment variable
    DB_NAME: "users_test"
    TIMEOUT: "30s"

If a referenced shell variable is not set, the literal string ${API_BASE_URL} is used as the value. Set a default with ${API_BASE_URL:http://localhost:5016}.

Default values

env:
  API_URL: "${API_BASE_URL:http://localhost:5016}/v1"

If API_BASE_URL is not set in the environment, API_URL becomes http://localhost:5016/v1.

2. System variables

These are always available without any declaration.

Variable	Description	Example value
`${FLOW_ID}`	Current flow ID (UUID)	`"abc123..."`
`${FLOW_NAME}`	Current flow name	`"User Registration"`
`${EXECUTION_ID}`	Current execution ID (UUID)	`"def456..."`
`${TIMESTAMP}`	ISO 8601 UTC timestamp	`"2026-03-16T10:00:00Z"`
`${TIMESTAMP_UNIX}`	Unix timestamp (seconds)	`"1742119200"`
`${RANDOM_ID}`	Random UUID	`"7f3a1b2c..."`
`${RANDOM_INT}`	Random integer 0–999999	`"482910"`
`${RANDOM_STRING}`	Random 16-char alphanumeric	`"k9mXpQ3rLtN8vWzY"`
`${FLOW_STATUS}`	In teardown: `"passed"` or `"failed"`	`"failed"`

${RANDOM_ID} and ${RANDOM_INT} are useful for generating unique test data that does not collide across concurrent runs.

- id: create_user
  action: http_request
  config:
    method: POST
    url: "${API_URL}/users"
    body:
      email: "test-${RANDOM_ID}@example.com"
      reference: "REF-${RANDOM_INT}"

3. Faker variables

Faker variables generate realistic test data on demand. They are evaluated fresh for each step execution.

body:
  first_name: "${FAKER.name.firstName}"
  last_name: "${FAKER.name.lastName}"
  email: "${FAKER.internet.email}"
  phone: "${FAKER.phone.phoneNumber}"
  street: "${FAKER.address.streetAddress}"
  city: "${FAKER.address.city}"
  company: "${FAKER.company.companyName}"
  product: "${FAKER.commerce.productName}"
  price: "${FAKER.commerce.price}"
  user_id: "${FAKER.datatype.uuid}"

Common Faker namespaces:

Namespace	Examples
`FAKER.name`	`firstName`, `lastName`, `fullName`
`FAKER.internet`	`email`, `userName`, `url`
`FAKER.phone`	`phoneNumber`
`FAKER.address`	`streetAddress`, `city`, `state`, `country`, `zipCode`
`FAKER.company`	`companyName`, `catchPhrase`
`FAKER.commerce`	`productName`, `price`, `department`
`FAKER.datatype`	`uuid`, `boolean`, `number`

Output path expressions

The right-hand side of each output entry is a path expression into the step result.

output:
  status_code: "response.status"
  user_id: "response.body.id"
  first_user: "response.body.users[0]"
  first_email: "response.body.users[0].email"
  all_headers: "response.headers"
  content_type: "response.headers[\"Content-Type\"]"
  duration: "response.time"

output:
  rows: "result.rows"
  row_count: "result.count"
  first_row: "result.rows[0]"
  user_email: "result.rows[0].email"
  user_status: "result.rows[0].status"

output:
  messages: "result.messages"
  message_count: "result.count"
  first_value: "result.messages[0].value"
  event_type: "result.messages[0].value.event_type"

JSONPath syntax

JSONPath can be used in output expressions and assertions for complex data access.

# Object property
"$.user.name"
"response.body.data.id"

# Array access
"$.users[0]"          # First element
"$.users[-1]"         # Last element
"$.users[0:3]"        # Elements 0, 1, 2 (slice)

# Filter
"$.users[?(@.status == 'active')]"          # Active users only
"$.products[?(@.price < 100)]"              # Products under $100

# Recursive descent
"$..email"                                  # All email fields at any depth

# Wildcard
"$.users[*].id"                             # All user IDs

# Functions
"$.users.length"                            # Array length
"$.prices.sum()"                            # Sum of numeric array
"$.prices.avg()"                            # Average
"$.prices.min()"                            # Minimum value
"$.prices.max()"                            # Maximum value

Variable interpolation features

Nested variable access

url: "${API_URL}/users/${user_id}/posts/${post_id}"

Conditional expression

message: "${status == 200 ? 'Success' : 'Failed'}"

Arithmetic

next_page: "${current_page + 1}"

String functions

uppercase_name: "${name.upper()}"
lowercase_email: "${email.lower()}"

Concatenation

full_name: "${first_name} ${last_name}"

All variables share a single flat scope for the duration of a flow execution. There is no block scoping. Variables set in setup steps are available in steps and teardown. Variables set in a for_each loop are available to subsequent outer steps after the loop completes.

If two steps define an output with the same variable name, the later step's value overwrites the earlier one. Use distinct, descriptive output names to avoid conflicts.