molted experiments

Command	Description
`molted experiments create`	Create a new experiment
`molted experiments list`	List experiments
`molted experiments get`	Get an experiment by ID
`molted experiments start`	Start an experiment (only from draft)
`molted experiments stop`	Stop an experiment (only from running)
`molted experiments results`	Get experiment results by variant

Agent context

Experiments compare template variants (or a holdout) on a journey step. Reachable with a Bearer API key (despite older docs implying cookie-only).

An experiment only actually runs when a journey step's stepId matches its journeyStepId exactly (journey-step-processor.ts) -- --journey-step is optional on the API, but an experiment with none attached to a real step never fires. Pass real journey/step UUIDs, e.g. from molted journeys get.

Create an experiment

molted experiments create --name "Subject line test" \
  --journey 3fa2c1e0-8b8e-4c1a-9e5a-6f6e2b6e6a11 \
  --journey-step 7c9c0d3a-2f5b-4a49-9b2e-1d3f9a7c6e02 --type ab \
  --variants '[
    {"id":"v1","name":"Control","weight":50,"templateVersionId":"tv_1","isControl":true,"isHoldout":false},
    {"id":"v2","name":"Casual","weight":50,"templateVersionId":"tv_2","isControl":false,"isHoldout":false}
  ]'
# → --variants also accepts @path/to/variants.json

Variant fields: id, name, weight (positive number, relative), templateVersionId (or null), subjectOverride (or null), timingDelayMinutes (or null; stored but not applied to sends yet), isHoldout (the contact gets no send), isControl (the baseline that results compare the other variants against). Note: "templateId" is NOT a variant field -- the API ignores it, so the CLI rejects it; use "templateVersionId" (from templates create/add-version).

Caveat (#1490): a variant's templateVersionId is stored and recorded as the send's templateId, but the send is still rendered from the journey step's own base template -- it is not re-rendered from the variant's template version. subjectOverride DOES take effect. Until #1490 ships, use subjectOverride for real A/B content differences.

Run it

molted experiments start <id>   # only from draft
molted experiments stop <id>    # only from running
molted experiments results <id>

List / inspect

molted experiments list [--journey <journeyId>]
molted experiments get <id>

Commands

molted experiments create

Create a new experiment

molted experiments create [options]

Option	Description	Default
`--name <name>`	Experiment name (required) (required)
`--variants <json-or-@file>`	JSON array of variants, or @path/to/file.json (required) (required)
`--journey <id>`	Journey ID
`--journey-step <id>`	Journey step ID -- required in practice: an experiment only runs when this matches the step being processed
`--type <type>`	Experiment type (server default: ab) (one of: ab, holdout)
`--segment <id>`	Segment ID to scope the experiment to

molted experiments list

List experiments

molted experiments list [options]

Option	Description	Default
`--journey <id>`	Filter by journey ID

molted experiments get

Get an experiment by ID

molted experiments get <id>

Argument	Description	Required
`<id>`		yes

molted experiments start

Start an experiment (only from draft)

molted experiments start <id>

Argument	Description	Required
`<id>`		yes

molted experiments stop

Stop an experiment (only from running)

molted experiments stop <id>

Argument	Description	Required
`<id>`		yes

molted experiments results

Get experiment results by variant

molted experiments results <id>

Argument	Description	Required
`<id>`		yes

See CLI overview for global options, credentials and output format.

On this page