Specialty nodes
Wait, Condition, Notification, and Audio nodes for timing, branching, email, and sound.
Once your flow has its core nodes and assertions, these nodes handle timing, branching, communication, and audio.
01Wait
The Wait node pauses the run for a set duration. Set the delay in minutes, hours, or days in the Configure panel. New Wait steps default to 30 minutes — review the delay when you add one so the flow resumes when you intend.
A Wait creates a segmented flow: execution pauses, then resumes automatically after the delay. Because the browser session ends during the pause, add a Login node after the Wait to re-establish context before later steps run. Step outputs from before the Wait remain available after the run resumes.
02Condition
The Condition node branches the flow on a runtime condition. Define what to evaluate, then connect separate paths for the true and false outcomes.
Cofactor evaluates a condition two ways:
- Agentic (default) — describe the condition in plain language and the agent evaluates it against the live page.
- Deterministic — compare a variable or the current user (Acting As) against a value using operators like equals, contains, is set, or greater/less than. Use this when you want an exact, repeatable branch with no agent judgment.
Condition fields support $ variable and macro insertion.
03Notification
The Notification node sends an outbound email when the run reaches it — to notify a team, share run context, or deliver files. Add it from the quick-create menu or by clicking an edge to insert it between two steps. Place it after the steps that produce the information you want to include.
Configure each Notification step:
| Field | Description |
|---|---|
| Recipients | One or more email addresses |
| Subject | The subject line |
| Body | The message content |
| Attachments | Optional files to send; filenames can use macros |
Recipients, Subject, Body, and attachment filenames all support variables and macros, so one saved step can adapt to each run or branch — route to different recipients, fold in run-specific details, or chain values from earlier steps.
The email sends only if the run reaches the node. If branching, an earlier failure, or another path skips it, no email goes out. A Notification step's outputs (its resolved recipients, subject, and body) are available to later steps, including after a Wait and resume.
Keep each Notification focused on a single purpose, write short subjects and bodies, and place the step after the state you want to communicate.
04Audio
Audio nodes test browser experiences that play or capture sound — voice prompts, call flows, microphone-driven onboarding.
Audio input
The Audio input node plays an uploaded or recorded clip into the browser microphone during a run. Place it after the step that opens the experience listening for input. You can loop the clip and choose whether the flow waits for playback to finish before continuing. Reuse the same clip across runs for repeatable voice tests.
Audio assertion
The Audio assertion node validates sound produced during the run. Add it where the audio you want to check should already be playing. It supports two styles of check:
| Check | What it verifies | Best for |
|---|---|---|
| Quantitative | Measurable properties — minimum/maximum duration, silence level | Confirming audio exists and isn't silent or truncated |
| Natural-language | Whether the audio communicates the expected meaning | Greetings, disclaimers, prompts, scenario-specific spoken details |
Start with a duration or silence check to confirm audio is present, then add a natural-language prompt for content. Allow margin in thresholds for natural pauses so expected silence doesn't fail the check. Combine with scenarios to validate different spoken output across runs. Outcomes appear on the step in run details.