How to Record Guides and Use Voice Commands in Glitter AI

Learn how to capture your screen, record voiceovers, and use voice commands to seamlessly edit your guides in real-time with the Glitter AI desktop app.

May 29, 2026 By Chris

This guide demonstrates how to capture screen activity and use real-time voice commands to edit your recordings using Glitter AI. By the end of this process, you will be able to dictate instructions, navigate through your required applications, and instruct the AI to backtrack or crop specific segments of your recording without manual video editing.

This process is ideal for technical support teams, onboarding specialists, and client-facing roles who need to generate quick software tutorials. It is especially useful when you want to simultaneously walk a client through a process on a live call while generating documentation in the background.

To begin creating documentation, you must choose how you want to capture your process.

Step #1: Click Download & Start Recording to install the Glitter AI desktop application. — Click **Download & Start Recording** to install the Glitter AI desktop application.

You can use Glitter AI while on a live call with a client. The application will run in the background, listening to your conversation and capturing your screen as you naturally walk the client through the process.

Recording Your Workflow

Once the recording begins, simply click through your standard workflow. The desktop application will quietly capture your actions while listening to your voice.

Step #2: Navigate through your required web applications, such as AI troubleshooting tools or search engines. — Navigate through your required web applications, such as AI troubleshooting tools or search engines.

Step #3: Update your tracking spreadsheets or project management documents as usual. — Update your tracking spreadsheets or project management documents as usual.

Step #4: Access any necessary web portals, such as security dashboards, to continue your demonstration. — Access any necessary web portals, such as security dashboards, to continue your demonstration.

Step #5: Review system statuses or agent health in your respective administrative consoles. — Review system statuses or agent health in your respective administrative consoles.

Step #6: If an unrelated system update or pop-up interrupts your process, click Remind me later to dismiss it without affecting the guide. — If an unrelated system update or pop-up interrupts your process, click **Remind me later** to dismiss it without affecting the guide.

Using Voice Commands

Glitter AI interprets your natural speech to format the guide. You can actively issue voice commands to edit the recording in real time.

If you make a mistake, simply tell the AI to fix it out loud. For example, say: "Glitter, backtrack everything I've said for the last 45 seconds, don't include that in the tutorial, and start the new screenshots at second 30." The AI will process this instruction and automatically crop out the error.

Managing the Recording

When you are ready to conclude or pause your guide, use the desktop widget overlay.

Step #7: Open the Glitter Desktop application window and click Stop. — Open the Glitter Desktop application window and click **Stop**.

Step #8: Use the widget controls to manage the current recording session. — Use the widget controls to manage the current recording session.

Control	Action
Stop	Ends the capture process and begins generating the final written guide.
Pause	Temporarily suspends the recording so you can prepare the next step.
Discard	Cancels the current session and permanently deletes the captured data without saving.

Q: Can I use this software while on a phone call with a client?

A: Yes, the tool can run in the background and record your steps while you talk the client through a process.

Q: Does the tool convert my speech to text?

A: Yes, it actively listens to your voice and transcribes it into coherent written instructions.

Q: How does the AI handle slang or unstructured speech?

A: The AI interprets and coordinates your language to create professional steps, ignoring conversational filler and translating slang into standard instructions.

Q: How do I fix a mistake made during the recording?

A: You can use natural voice commands, such as explicitly telling the app to "backtrack," to instruct the AI to ignore the last few seconds or minutes of your recording.

Term	Definition
SentinelOne	An endpoint security platform used for threat detection and response.
Huntress	A managed security platform focused on detecting hidden threats and vulnerabilities.
Wireshark	A network protocol analyzer used for capturing and examining data traffic.