> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bowerlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Voice commands by device

> What you can say to Bird hands-free, and how it differs across mobile, Meta Ray-Ban glasses, and VITURE Luma Ultra AR glasses.

Bird's [live agent mode](/ai-and-search/live-agent-mode) is voice-driven, but **what you can ask for depends on the device's capabilities**. A phone has a touchscreen, the Meta Ray-Bans have no screen at all, and the VITURE Luma Ultra glasses put a heads-up display and your hands into the mix. This page is the quick reference for which commands work where.

You don't need to memorise exact phrases — Bird understands intent, so paraphrases of the examples below work too. The phrases are just reliable starting points.

## At a glance

| What you want to do                            | Mobile / Desktop |          Meta Ray-Ban          |  VITURE Luma Ultra |
| ---------------------------------------------- | :--------------: | :----------------------------: | :----------------: |
| Run a protocol & step through it by voice      |         ✓        |                ✓               |          ✓         |
| Ask questions, talk through your work          |         ✓        |                ✓               |          ✓         |
| Start / stop recording a clip                  |         ✓        | ✓ (REC shown in the phone app) |          ✓         |
| Take a snapshot / create a note                |         ✓        |     ✓ (spoken confirmation)    |          ✓         |
| **Show** an image or reference video on screen |    ✓ (overlay)   |          — (no screen)         | ✓ (in-world panel) |
| Jump a reference video to a step               |         ✓        |                —               |          ✓         |
| Reposition / resize the on-screen display      |         —        |                —               |          ✓         |
| Put the display to sleep / quit hands-free     |         —        |                —               |          ✓         |

<Note>
  The differences are about **capability, not preference** — the Ray-Bans have no screen, so "show me the gel photo" has nowhere to appear. Commands that need a display only exist on devices that have one.
</Note>

<Note>
  **Jumping to a specific step is confirmed first.** On any device, when you ask to jump to a named or numbered step — *"skip to step 3"*, *"go to the vortex step"* — Bird reads the step's title back (*"Jump to 'Vortex the sample' — that right?"*) and waits for a *"yes"* before moving. Marking a step done, skipping, and going back stay instant. See [execution mode](/ai-and-search/live-agent-mode#execution-mode-running-a-protocol) for details.
</Note>

## Commands by device

<Tabs>
  <Tab title="VITURE Luma Ultra">
    The richest surface: a heads-up protocol card, an in-world image/video panel, hand gestures, and voice. Everything below is hands-free.

    **Run and step through a protocol**

    * *"Run the DNA extraction protocol"* / *"Start the execution"*
    * *"Next step"* · *"Mark this done"* · *"Go back"* · *"Repeat that"*
    * *"Skip to step 3"* · *"Show me the vortex step"*

    **Show a reference video or image**

    * *"Show me the video for this protocol"* — plays the bound step-video in the panel
    * *"Show me the gel photo"* (or any saved image)
    * *"Skip to step 2"* — seeks the video to that step's chapter

    **Control the media panel** (once it's showing)

    * Pause / resume: *"Pause the video"* · *"Keep playing"*
    * *"Restart this step's video"*
    * Close it: *"Close the video"* · *"Stop the video"* · *"That's enough"* · *"Get rid of it"*
    * Or **reach out and hold your fingertip on the × button** in the corner of the panel
    * Or **pinch with both hands and move them together / apart** to resize it

    **Move the heads-up card**

    * *"Pin it here"* (leave it fixed in space) · *"Follow me"* (re-dock to your view)
    * *"Move it up"* · *"Bring it closer"* · *"Move it left/right/down/farther"*
    * Or **pinch with both hands** to resize the card

    **Sleep, record, and quit**

    * *"Go to standby"* / *"Wake up"* (or **tap the standby pill** to wake)
    * *"Start recording"* / *"Stop recording"*
    * *"Quit Bower"* — the glasses ask you to **confirm on-screen** before exiting

    <Tip>
      If a command that controls the display ("close the video", "pin it here") ever isn't understood, make sure you said the *display* word — e.g. "close the **video**" rather than just "stop". Bird needs to know you mean the on-screen panel, not the recording or the protocol.
    </Tip>
  </Tab>

  <Tab title="Meta Ray-Ban">
    No screen, but the camera and mic are live, so replies are spoken and Bird keeps them short.

    **Run and step through a protocol**

    * *"Run the buffer prep protocol"* / *"Start it"*
    * *"Next step"* · *"Mark this done"* · *"Go back"* · *"What's the next step?"*
    * *"Skip to step 3"* · *"Read me step 5"*

    **Capture by voice** — the glasses have a camera, so this works even without a screen

    * *"Take a snapshot"* / *"Save this"* — Bird grabs the current camera frame and saves it to your workspace
    * *"Make a note that the sample looked cloudy"* — Bird records a note by voice
    * Bird confirms out loud; there's no screen to preview the result, but the snapshot or note lands in your workspace just the same

    **Record a clip**

    * *"Start recording"* / *"Record this"* · *"Stop recording"* / *"That's enough, you can stop"*
    * Bird confirms out loud, and the clip lands in your workspace the same way it does on any other device ([details](/ai-and-search/live-agent-mode#recording-video-clips))
    * The glasses have no screen, so the red **REC** indicator appears in the Bower app on your paired phone for as long as the recording is active

    **Talk through your work**

    * Ask anything: *"How much sodium chloride does this step need?"*, *"Is 4 °C okay for this?"*

    **Not available on Ray-Ban** (needs a screen): *showing* images or videos, the heads-up card, resizing/pinning a display, and standby. Use mobile or the VITURE glasses for anything that has to appear on screen.
  </Tab>

  <Tab title="Mobile / Desktop">
    Live mode in Bird chat — voice plus an on-screen overlay and your camera. See [Live agent mode](/ai-and-search/live-agent-mode) for the full walkthrough.

    **Run and step through a protocol**

    * *"Run the PCR setup protocol"* · *"Next step"* · *"Mark this done"* · *"Skip to step 3"*

    **Show media on screen**

    * *"Show me the gel from last week"* (or any saved image) — appears as an overlay
    * *"Show me the video for this protocol"* · *"Skip to step 2"* to seek it
    * Close the overlay by **tapping the ×** (or tapping outside it)

    **Capture and record**

    * *"Take a snapshot"* — saves the current camera frame
    * *"Create a note from what we just discussed"*
    * *"Start recording"* / *"Stop recording"* — saves a short clip ([details](/ai-and-search/live-agent-mode))

    **Video awareness**

    * With video on, point your camera and ask: *"What's the reading on that display?"*
  </Tab>
</Tabs>

## Related

* [Live agent mode](/ai-and-search/live-agent-mode) — how to start a live voice session
* [Protocol step videos](/ai-and-search/protocol-step-videos) — binding reference videos to protocol steps
* [Bird AI assistant](/ai-and-search/bird-ai-assistant) — the assistant behind all of this