All keyboard shortcuts for AI for video

Shortcuts for AI video: Runway, Descript, and CapCut

With the right keyboard shortcuts, you can access AI features faster and work more efficiently on generated video, text-to-speech, and automatic editing. This guide covers Runway, Descript, and CapCut. You’ll see which keys directly access an AI feature, which open a command menu, and which do not have a separate standard shortcut assigned.

An important distinction: a program can have dozens of AI features without offering dozens of AI keyboard shortcuts. Generating a video, replacing a background, and recognizing subtitles are often done via a button or a command field.

The genuine documentation consulted does not list a hundred different, AI-specific keyboard shortcut combinations for these three applications. Therefore, this overview provides the substantiated controls. Platform variations and different commands within the same menu are not presented as a hundred unique keyboard shortcuts.

How to Read the Keys and Instructions

• Ctrl: the Control key on Windows.
• Cmd: the Command key on a Mac, identified by ⌘.
• Shift: the key you also use to type uppercase letters.
• Enter or Return: the confirmation key.
• A plus sign: Hold down the listed keys simultaneously.
• Next: Do not perform the next step until the previous one is complete.
• Focus: the element currently receiving input from your keyboard.

To press Ctrl+K, hold down Ctrl while pressing K. When you “open the command menu and search for Regenerate,” you’re performing a sequence of multiple steps.

A single letter can also be a shortcut key. Its function depends heavily on the active mode. When typing text, a letter typically enters text rather than opening a tool.

Overview: Confirmed access to AI features

No.	Application	Windows	Mac	Feature	Important context
1	Runway Agent	/	/	Opens the selection for Agent Skills.	Type the slash into the Agent input field.
2	Descript	D	D	Launch Regenerate.	Use the appropriate script context; single letters do not work as keyboard shortcuts in Write Mode or Correct Mode.
3	Descript	Ctrl+K	Cmd+K	Open the Action bar to search for commands and available AI features.	This is general access; it is not a direct instruction to generate a video.

These are three confirmed access procedures, including platform variants. They do not represent three fully automated video productions. After opening the file, you’ll usually need to select a function, enter data, or check settings.

Runway: Create and Product Type AI Videos

Using the Slash in Runway Agent

Runway describes Agent Skills as guided workflows for creative tasks. To access the options, type / in the Agent input field. Then select an available skill and add information about your desired outcome.

The same input can trigger different types of offers. The selected process and your instructions determine what Agent does next.

Described Agent Skill	What is it for?	What do you provide?
Commercial	Developing a commercial.	The product, the target audience, the main message, and the desired closing line.
Motion Graphic	Convert text, charts, or infographics into moving images.	The content, desired style, and order.
Ad Campaign	Develop a campaign using video and visual content.	The product range, visual direction, and desired formats.
Mood Board	Exploring a visual direction.	Colors, atmosphere, themes, and examples of the desired look and feel.

These examples are options under the same access. “Commercial” and “Motion Graphic” are therefore not separate key combinations.

A clear brief for a short commercial

Start with a single message. A video that simultaneously explains a product, lists all its benefits, and introduces the entire company can quickly become overwhelming.

A sample task to complete after selecting a suitable skill:

“First, develop a concept for a short vertical product video. The target audience consists of remote workers. The main message is that an organized workspace promotes peace of mind. Use three clear shots and conclude with space for a brief call to action. Show the layout first.”

This is a sample prompt; it does not guarantee that every desired feature will be implemented exactly as described. Please review the suggested scenes before proceeding.

Pay special attention to the flow: does the first shot clearly establish what the video is about? Does the second shot build on that? Does the conclusion make sense without needing further explanation?

Differences between text-to-video and image-to-video

In its prompting guides, Runway distinguishes between Text to Video and Image to Video. With Text to Video, you describe both the visible scene and the movement. With Image to Video, the image provides the visual basis, and the prompt focuses primarily on what needs to move.

How It Works	What are you focusing on in your description?	Example
Text to Video	Subject, setting, action, and camera movement.	A cyclist rides along a quiet canal while the camera moves parallel to them.
Image to Video	Motion and development based on the provided image.	The camera slowly moves forward; leaves sway gently in the wind.

Don’t use the image merely as a decorative example when it’s supposed to serve as the starting image. Consider in advance what role your input plays and verify that you’ve chosen the correct generation method.

Describing a camera movement in concrete terms

“Make it cinematic” provides little information about the desired shot. Instead, describe how the camera moves and what should remain in the frame.

“A calm, uninterrupted shot. The camera slowly moves toward the product. The product remains fully visible and does not change shape. The background remains calm.”

Next, evaluate the entire clip. A good opening shot doesn’t necessarily mean the movement halfway through is also correct. Look for changing shapes, shifting details, and movements that suddenly accelerate.

Editing an existing video

Runway describes Edit Studio as a tool for editing existing recordings using text-based instructions. Examples include changing a background, removing unwanted objects, and adjusting the overall look and feel.

A targeted task is easier to monitor than a general improvement:

“Replace the busy background with a quiet office setting. Keep the person, clothing, posture, and genuine camera movement.”

Check whether the desired change remains consistent throughout the entire recording. Also pay attention to edges, reflections, and parts that temporarily disappear behind the subject.

No confirmed direct shortcut for every generation

The Runway manuals consulted describe the “Generate” button for image and video generation. They do not specify a universal keyboard shortcut that performs the same function in every generation mode.

Therefore, do not present the Enter key out of context as “the shortcut to generate a Runway video.” Similarly, the slash (/) is specific to Agent Skills.

Description: AI editing via Regenerate and the Action Bar

Open Regenerate with D

Descript’s genuine list of keyboard shortcuts lists “D” for Regenerate. This feature uses AI to enhance or recreate parts of a recording.

Active mode is important. The description states that individual letters function as shortcut keys in the Script Editor when Write Mode and Correct Mode are not active.

First, select the desired component, then check which Regenerate option is available. Enabling this feature does not mean that every recording can be processed in the same way.

What can Regenerate change?

Descript describes applications for improving awkward editing transitions, adjusting spoken words, and smoothing out audio and video. The current manual mentions support for English-language audio and video.

When spoken words are replaced, the described process uses an authorized, customized AI voice. Additional restrictions apply to video generation, depending on the recording.

So be sure to check the language and recording compatibility first. Don’t automatically assume that the same features are available just because a video is in Dutch.

Open the Action Bar

Press Ctrl+K on Windows or Cmd+K on a Mac to open the Action Bar. Use this general command bar to search for available features.

Goal	Directions	What does this route do?
Visit Regenerate	Open the Action bar and search for Regenerate.	Searches for the command to open the Regenerate interface.
Review detected stop words	Open the Action bar and search for “Remove filler words.”	Opens the list of detected stop words.

These are two uses of the same access. They are not two unique AI keyboard shortcut combinations. Check the result you found before executing a command.

Carefully review keywords

An “uh” may be superfluous, but it can also be part of a natural hesitation. Therefore, do not remove it solely based on the detected text.

Listen to the sentence before and after the deletion. Does the transition sound natural? Is there enough breathing room? Does the rhythm match the speaker’s style?

When watching a video, also check the visuals. A smooth sentence can coincide with a visible jump in head position or hand movement. Audio and visuals therefore each require their own assessment.

Correcting a transcript is different from replacing speech

A misrecognized word in the transcript doesn’t necessarily mean the recording is incorrect. If someone says the correct word, a text correction may be sufficient.

If you want to change what is actually being said, that involves a different process. Determine in advance whether you’re simply correcting the written text or creating brand new spoken content.

For example: a misspelled product name in the subtitles does not automatically require the voice to be regenerated.

Access Underlord via the selection

"Descript" describes "Ask Underlord" in the selection toolbar. This allows you to start a conversation with the AI co-editor based on the selected content. The documentation consulted does not specify a separate universal keyboard shortcut for this.

An example of how to define a scope of work:

“Review this selection for repetition. Suggest which phrases can be omitted, but retain all terms, product names, and numbers. First, briefly explain what you would remove.”

This is an example of a targeted instruction. Check which action the available interface offers and evaluate the proposed change for its significance.

CapCut: AI features for video and product type

CapCut features automatic subtitling, text-to-speech, automatic reframing, and background removal, among other things. For the features listed below, the manuals consulted do not specify a separate universal keyboard shortcut.

Overview of features and access methods

AI feature	What do you use these for?	Access Description	Have you set up your own custom keyboard shortcut?
Auto Captions	Convert speech to subtitles.	Open Captions, select Auto Captions, check the language and source, and start recognition.	No.
Auto Reframe	Resize the image to a different format.	Use the relevant feature among the available video editing options.	No.
Auto Cutout	Automatically separate an object from the background.	Select the clip and use the appropriate cut function.	No.
Script to Video	Moving from a text idea to a video proposal.	Open the available Script to Video interface and follow the steps.	No.

The exact screen layout may vary between desktop, web, and mobile. Therefore, use the feature name as a reference point and check which options your version displays.

Generate automatic subtitles

CapCut describes the process for desktop via Captions and Auto Captions. Select the correct audio source and language, then choose Generate. After that, you can review the text, timing, and appearance.

Good fact-checking starts with names, numbers, and short denials. The difference between “works” and “doesn’t work” changes the entire message.

Then read the subtitles as you listen. Check to make sure each line appears when the words are spoken and stays on screen long enough to read.

Improving recognition

CapCut lists unclear audio, background noise, incorrect language settings, and connection issues as possible causes of failed recognition.

First, check whether the intended voice is clearly audible and whether you’ve selected the correct language. Restarting automatic recognition won’t help much if the input continues to contain the same problem.

If you find an error in a single word, don’t immediately redo the entire subtitle. Consider whether a manual correction is simpler and preserves existing improvements.

Review automatic reframing

When switching from landscape to portrait orientation, important information may fall outside the frame. Examples include a second person, a product in someone’s hand, or a clue at the edge.

So don’t just check whether the main plot is being followed. Also make sure the viewer has everything they need to understand the action.

In a demonstration, a close-up focused on the face may be less effective than one that keeps the hands and product visible.

Background removal in moving images

A shot that looks good in a single frame may appear choppy when in motion. Check for hair, fingers, thin objects, and fast movements.

Also pay attention to the moment when an object moves in front of or behind the subject. That’s often where you can most clearly see whether the separation between the subject and the background remains stable.

Place the result on the intended background. A light border may be barely visible on white but will stand out immediately on a dark colour.

A reliable approach to AI video

1. Set one goal per section

Write down what the viewer should understand or feel after watching the clip. A product introduction requires different visuals than an instructional or atmospheric clip.

Use that goal as a selection criterion. A generation that stands out but contributes little to the message may not be the best choice.

2. Start by writing a brief description of the shot

For each shot, describe the subject, the action, the camera, and the desired endpoint. This will help you determine whether the movement seems feasible within the chosen duration.

A simple description might be, for example: “The camera moves toward the desk; the hand places the product down; the product remains clearly visible.”

3. Change one key component at a time

When working on a follow-up version, adjust either the movement or the composition first. If you change the environment, style, action, and camera all at once, it becomes difficult to compare.

Make a brief note whenever you take a successful step in the right direction. For example, note which approach worked and which details still need improvement.

4. Check the total duration

Don’t just judge by the preview image. Watch the beginning, middle, and end, and pay attention to changes in shape, sudden objects, and unexpected transitions.

Also, check whether the very last second is usable for your edit. Even a strong clip can be difficult to transition into if it ends with a sudden movement.

5. Evaluate audio and video separately

First, listen to the spoken content, and then focus on the visuals. That way, you’ll notice an unnatural pause, strange mouth movements, or a visible jump sooner.

Be sure to double-check names and numbers in written form as well. Even a pleasant-sounding voice can still mispronounce a product name.

6. View the result in the final product type

A snippet can work well on its own but still be too busy when placed behind a title. Therefore, add text and other necessary elements before making your final choice.

Make sure the video is viewable on a small screen as well. Subtitles, product details, and key offers must remain clear and understandable on a small screen.

Why does an AI shortcut sometimes not work?

• The focus is incorrect: you’re typing in a text field when you’re expecting a tool.
• Another mode is active: Descript disables individual letters as shortcut keys in Write Mode and Correct Mode.
• You’re using a different environment: Agent, a generation tool, and a timeline editor all have different controls.
• The combination has been adjusted: an example from a video may use a personal setting.
• This feature is not available: check the options visible in your version and account.

Do not repeatedly press a key combination if you cannot see which item is selected. First, check the active environment and search for the desired function by name.

Frequently Asked Questions About Keyboard Shortcuts for AI Video

Are these 100 AI keyboard shortcuts?

No. The genuine documentation does not list a hundred different direct combinations for this classification. This reference work contains the substantiated AI approach and practical explanations, without counting standard editing functions as AI commands.

Which keyboard shortcut opens Regenerate in Descript?

Descript mentions D. Use the correct context in the Script Editor and verify that Write Mode and Correct Mode are not active.

Which key combination opens the command menu in Descript?

Press Ctrl+K on Windows or Cmd+K on a Mac. Then search for the desired available command. This shortcut opens the Action bar but does not generate a video on its own.

What does the slash in “Runway” mean?

In the Agent input field, pressing the slash key opens the Agent Skills menu. Then select a skill and add information about your assignment.

Does CapCut have a confirmed universal keyboard shortcut for Auto Captions?

The genuine manuals consulted describe a method using Captions and Auto Captions. They do not specify a single universal keyboard shortcut.

Why aren’t “Cut” and “Play” included in this list?

Because they don’t execute AI commands themselves. They’re useful for video editing, but fall outside the scope of a list that promises only AI shortcuts.

Which operating system should I learn first?

Learn how to access the AI feature you’re using, the required selection, and how to evaluate the results. Afterward, only add combinations whose context and functionality you understand.