Run Mode

Run Mode determines when actions execute after a phrase matches.

Action Phrase can recognize speech continuously — but you control whether actions run automatically, require confirmation, or depend on an activation phrase.

Run Mode is about safety, speed, and intent.

Available Run Modes

Auto

Actions run immediately when a phrase matches.

Best for:

• Rehearsed environments
• Predictable commands
• Confident, deliberate speech

Auto mode is the fastest workflow. When the phrase matches, its actions execute instantly.

Manual

When a phrase matches, actions do not run automatically.

Instead:

• The matched phrase is displayed.
• You must confirm before actions execute.

Best for:

• High-risk environments
• Testing and setup
• Situations where accidental triggers must be avoided

Manual mode prioritizes safety over speed.

Smart

Smart mode combines automation with intentional activation.

In this mode:

• Actions run automatically only when the phrase is preceded by your Activation Phrase.
• If the activation phrase is not spoken, manual confirmation is required.

This gives you the flexibility of Auto mode — with the control of Manual mode.

Example:

Activation Phrase:

“Computer”

You say:

“Computer, show lower thirds.”

→ Actions run automatically.

You say:

“Show lower thirds.”

→ The phrase matches, but requires manual confirmation.

Smart mode is ideal for live environments where you want natural speech without constant manual approval.

Command Boundary Modes

Automatic (Silence Detection)

Action Phrase listens continuously and decides command boundaries from speech flow. It starts matching as soon as you speak and runs the phrase after trailing silence is detected. This is the default mode and works well for quick, natural commands.

Start Phrase

This mode waits for a specific Start Phrase before it captures text for matching. It helps cut down false triggers in noisy environments.\ Available with Core.

Start & End Phrases

With this mode, only the words spoken between your Start Phrase and End Phrase are considered for command matching. It’s ideal when you want strict, explicit control over exactly what gets interpreted.\ Available with Studio.

End Phrase Only

In this mode, Action Phrase listens continuously, then uses trailing transcript context and your configured End Phrase to extract and match the command right before the end marker. Use this when you prefer “speak, then close with a phrase” workflows.\ Available with Studio.

Press and Hold to Recognize

The Press and Hold to Recognize setting changes how listening behaves.

When enabled:

• Phrases are only recognized while you are actively pressing the floating Listen button.
• Releasing the button stops recognition immediately.
• No passive listening occurs.

When disabled:

• The app listens continuously.
• Phrases can match at any time while listening is active.

Why Use Press and Hold?

Press and Hold provides physical intent signaling.

It is useful when:

• You are in a noisy environment.
• You want zero accidental triggers.
• You prefer tactile control.
• You are speaking in a live space with background conversation.

It functions similarly to push-to-talk systems used in professional communication.