How Content Creators Are Revolutionizing Video Editing with AI Coding Tools

Content creators on Instagram and other platforms are increasingly demonstrating how to edit videos using AI-driven code tools such as Claude Code with the Opus 5.5 model and Codex with GPT-6.1 Sol. Instead of manually editing every detail in traditional software, these tools allow users to build editing workflows using natural language instructions and code.

Creators are offering downloadable skills—instruction files that teach the AI how to perform specific tasks, which working methods to use, what rules to follow, and which tools to activate. In the case of video editing, a skill can define a workflow that includes cutting clips, adding subtitles, adjusting video formats, integrating music, and exporting the file.

We asked Bar Shaltiel, an entrepreneur, artificial intelligence lecturer, and founder of Practi AI, to explain how he edits videos using AI and to share the skill he created for this purpose.

Setting Up the Editing Environment

Before getting started, users need a Windows 10/11 or Mac computer running macOS 15 or later, approximately 8 GB of free disk space, and an appropriate subscription (even a $20 per month subscription will suffice) to Claude Code or Codex. Installation requires an internet connection.

Video Studio is a skill—a package of instructions and tools that adds dedicated capabilities to the agent, specifically Hebrew video editing. It operates within the computer's workspace, meaning that simply opening a regular chat conversation is not enough.

After downloading the skill file, video-studio.zip, from the Practi AI club skill page, attach it to a conversation in the Code environment and request: Install the Video Studio skill from the attached file, and verify that the installation is complete. At the end, explain how to open a new conversation with the skill. After installation, open a new conversation and write: Install a video editing studio for me. The initial setup takes between 5 and 20 minutes.

Starting with Clean Cuts

Shaltiel recommends starting with basic editing and only adding subtitles and effects afterwards. This ensures that the message is clear and that the transitions between segments work well. Drag the videos into the conversation window and explain what you want to do with them. For example: Connect the four videos into one Instagram Reel, in chronological order. Cut out silences, stutters, and repetitions. At this stage, no subtitles and no effects—just clean cuts and smooth transitions between segments.

The tool transcribes the speech and suggests which parts to remove. Before approving, review the suggestion to ensure no important sentence was cut. If you said the same sentence twice, you can ask to keep the more successful attempt. There is no need to manually delete all mistakes from the raw footage beforehand.

Adding Subtitles and Choosing Styles

Once the cuts are approved, request a version with subtitles. You can choose subtle text, large words, or subtitles where the spoken word is highlighted during speech. For example: Add Hebrew subtitles, highlighting the spoken word and keywords. Show me the style sheet first so I can choose.

The tool displays eight styles overlaid on your footage. You can choose the desired style or even combine options: the subtitles of one style and the colors of another. Before generating the video, review the transcription and correct errors. Names of people, products, and English terms require special attention.

Enhancing the Message with Effects

Transitions between topics, a short zoom on an important sentence, or an emoji at an amusing moment can add rhythm. To achieve a precise result, specify what to add and where. For example: Add a transition sound between topics and a short zoom on important sentences. When the speaker says free, add confetti. Use effects in moderation.

The effects are timed according to the words in the transcription, so confetti appears when the word is spoken, and zooms emphasize the chosen sentence. If the result feels too cluttered, you can request a targeted modification: Remove the emojis and transition sounds. Keep only the zoom on the central sentence.

Translating Explanations into Animations

When presenting data, a list of tips, or a process, you can turn them into animations. The voice continues to play, and instead of seeing the speaker throughout the entire video, the viewer sees a visual representation of the explanation. For example: While I am explaining the three steps, display a step-by-step building list at the pace of speech. Keep my voice in the background, and return me to the screen at the end.

Turning Long Videos into Multiple Reels

A lecture, lesson, or podcast can be turned into several short videos, each focusing on a single idea. Attach the recording and define how many clips you want, their length, and where they are intended to go. For example: Find three clips in the recording suitable for 30-45 second Reels. Each clip must address a single idea and be understandable on its own. Propose the clips to me before editing.

Reviewing, Correcting, and Exporting

Before producing the full video, the tool displays images from the editing stage for approval. This is the time to check the placement of subtitles, colors, and animations. After production, watch the video from start to finish and check the cuts and sound.

Every change is saved as a new version, allowing you to continue making corrections using short instructions. Shaltiel's recommendation is to start with a short video and a simple request. Approve the cuts, add subtitles, and only then decide whether effects and animations are needed.