If you want to know how to create AI videos, or if you have footage you want to edit using AI, don’t rush to search for a bunch of tutorials just yet. There are now dozens of AI editing Skills (skill packs installed on AI, specifically for video production and editing), and you can’t tell which ones are good or bad just by looking at their names. I’ve gone through the entire list and narrowed it down to 12 AI editing Skill repositories I could find, each labeled with a star rating and URL. Below, I’ve organized the information not by function, but by “what kind of footage you have”—dividing it into three scenarios: nothing at all, video without audio, and video with audio that just needs editing. Just pick the section that applies to you—they’re all ready-to-use AI automatic editing solutions.
If you're not sure where to find these skills, you can start by checking out the list I compiled earlier. 8 Platforms Where You Can Find Agent Skills...Once you've gotten a handle on the repository sources, come back and read this article.
Scenario 1: You have nothing but a single video you can post
These AI-powered automatic video editing tools can take a topic and handle everything from the script to the visuals, voice-overs, and subtitles—all the way through to the finished video—even without any source material. They’re arguably the most convenient solution to the question of “how to make an AI video.” You don’t need to provide any source material; you just need to decide whether to publish it. The difference between the three projects lies in the “genre” they produce.
calesthio/OpenMontage calesthio/OpenMontage
★★★★★ 73.2k stars
https://github.com/calesthio/OpenMontage
It’s in a league of its own at the top of the rankings: the combined number of stars for second through eighth place doesn’t even match its total. An open-source AI video production system featuring 12 production pipelines, over 100 tools, and more than 700 skill documents—it can handle any genre. The trade-off is that it’s a hefty installation with numerous dependencies, and initial setup takes some time. If you’re serious about creating AI videos, starting here is the surest way to stay on track.
Vincentwei1021/anything2explainer Vincentwei1021/anything2explainer
★★★★☆ 2.9k stars
https://github.com/Vincentwei1021/anything2explainer
It only produces explanatory videos. Given a topic, it creates a black-background animated video with voiceover and subtitles, and automatically generates a rhythm bar. Since it doesn’t do anything else, every video has a consistent look, making it easier for viewers to recognize the series. It’s best suited for a science and education channel.
ArcReel arc-reel/ArcReel
★★★★☆ 3.8k stars
https://github.com/arc-reel/ArcReel
Focus exclusively on feature films. Start with a novel or screenplay and develop it through character design, storyboarding, and all the way to the finished film. The selling point is character consistency across shots—short series often suffer when the lead’s face changes in the second episode, and this approach specifically addresses that issue. If you have a story but no actors or locations, take this route.
Scenario 2: There’s video but no audio; the film must rely on subtitles and pacing to hold the audience’s attention.
The common feature of this section isYou don't need to say a word...and there’s no need for real human voices either. The visuals can be generated, screen-recorded, or computer-generated—the content is driven entirely by subtitles, animations, and music. This approach is ideal for “no-on-camera” content such as tutorials, product demos, and data animations.
Vincentwei1021/video-shotcraft Vincentwei1021/video-shotcraft
★★★★★ 15k stars
https://github.com/Vincentwei1021/video-shotcraft
124 shot recipe cards that turn the nitty-gritty details—like cut points and sound effect placements—into ready-to-use skills, specifically designed to ensure silent scenes don’t look like a PowerPoint presentation. Whether it’s a product promotional video or a product demo without voiceover, this is all you need. Render with Remotion, and you can even edit the final video.
remotion-dev/html-video remotion-dev/html-video
★★★★☆ 4.2k stars
https://github.com/remotion-dev/html-video
Generate MP4 files directly using HTML/CSS and data. Anyone who’s built websites will immediately understand how it works—every div animation you’ve ever created can be turned into individual frames. Whether it’s data visualization, animated reports, or feature demos, it’s incredibly efficient, and you can regenerate the output just by changing a number.
Scenario 3: Sound and video are in place—all that’s missing is the “edit”; leave it to the voiceover.
The biggest difference from the previous section:You recorded the footage yourself, and the content consists mainly of people speaking.—Voiceovers, interviews, screen-recording tutorials, and vlogs. This section actually covers what AI video editing truly is: the footage already exists, and you let the AI decide what to cut and what to keep. The three types of tools handle this differently: you tell it what to cut, let it decide on its own, or have it automatically cut based on the transcript.
You say the word, and it acts (command-based)
Agentchengfeng/chengfeng-video-editing-skills Agentchengfeng/chengfeng-video-editing-skills
★★★★☆ 3.2k stars
https://github.com/Agentchengfeng/chengfeng-videocut-skills
"Cut the beginning, change the background music"—just say the command and it does exactly that, without touching the timeline. This is the AI video editing method that most closely resembles "giving instructions to a person." It has the lowest learning curve, but at the cost of precision—it’s good enough for general edits, but it’s a bit of a stretch to make frame-by-frame adjustments.
0xsline/OpenChatCut 0xsline/OpenChatCut
★★★★☆ 2.8k stars
https://github.com/0xsline/OpenChatCut
It’s also dialogue-based, but retains a professional multi-track timeline and uses Remotion for rendering, with a focus on local processing. Once you’ve finished editing the dialogue, you can still switch back to manual adjustments if you want to tweak individual frames. This provides an extra layer of security compared to purely dialogue-based systems, though the trade-off is that the interface isn’t quite as user-friendly.
If you don't tell it, it analyzes the source material on its own (in a way that involves understanding).
FireRedTeam/FireRed-OpenStoryline FireRedTeam/FireRed-OpenStoryline
★★★★☆ 3.2k stars
https://github.com/FireRedTeam/FireRed-OpenStoryline
"Intent"-driven, directorial editing. You simply state what you want to convey, and it decides on its own how to arrange the shots and which clips to cut. When you have a pile of disorganized footage and only a vague story in mind, this is less hassle than giving instructions one by one. Conversely, when your requirements are extremely clear, it actually gets in the way.
Remove Unnecessary Text from Transcripts (Automatic)
watzng/claude-video-editor watzng/claude-video-editor
★★★☆☆ <100 stars
https://github.com/watzng/claude-video-editor
Whisper first transcribes your speech into text, then flags any slip-ups or pauses that need to be cut from the transcript—all you have to do is review it in your browser. Right now, the most reliable approach to AI-powered automatic editing is “understand first, then edit.” It works best for screen-recording tutorials and instructional videos because it recognizes phrases like “um,” “uh,” and “let me try that again.” With fewer than 100 stars, it’s a small, clean tool—just don’t expect it to handle heavy-duty tasks.
nateherkai/herk-edit-kit nateherkai/herk-edit-kit
★★★☆☆ 2.4k stars
https://github.com/nateherkai/herk-edit-kit
The approach is similar to the previous one, but this time we’re creating vertical short videos: 15 skills, 406 motion graphics cards, with the system automatically selecting shots and adding motion effects based on the script. Vertical voice-over videos and vlogs are the most hands-off—you can post them right after editing. If you don’t create vertical videos, you won’t get much use out of this app.
After the video is finished: transcode, trim, and resize
This team isn't involved in the creative process; they handle the final touches. You might not touch it for a week, but you'll need it every time you release an album.
kajisho5/ffmpeg-skill kajisho5/ffmpeg-skill
★★★★☆ 4.1k stars
https://github.com/kajisho5/ffmpeg-skill
42 FFmpeg tools come with a user guide that even a computer can understand. Slice videos, add subtitles, adjust aspect ratios, tweak volume, and batch export in various sizes—all right on your local machine, with no need to upload your files. Once you install it, you’ll forget it’s even there, but you’ll miss it the moment you delete it.
OpenClip-App/agent-skills OpenClip-App/agent-skills
★★★☆☆ <100 stars
https://github.com/OpenClip-App/agent-skills
Split long videos into short, vertical clips with subtitles, while also offering transcription, format conversion, thumbnails, and background removal. If you have a one-hour recorded video and want to extract a dozen or so short clips to post, this is the easiest way to do it.
Two things you can't do without, even if you don't install them
These two aren't optional; they serve as the foundation and supply stations for other tools.
remotion-dev/skills remotion-dev/skills
★★★★☆ 4.6k stars
https://github.com/remotion-dev/skills
Remotion's official Agent Skills handle animations, subtitles, rendering, maps, and multimedia. It's not a ready-made tool, but a foundation—several of the projects mentioned earlier (video-shotcraft, html-video, OpenChatCut) run on it. Building your own workflow is a prerequisite; it's not really a matter of choosing whether or not to use it.
skills-directory/Generative-Media-Skills skills-directory/Generative-Media-Skills
★★★☆☆ 3.9k stars
https://github.com/skills-directory/Generative-Media-Skills
A collection of multimodal skills for images, video, and audio. It’s not meant for direct production—it’s more like a repository of ingredients. Whenever you’re missing a specific capability, just dig in and find what you need to fill the gap.
How to Choose: First, see what you have on hand
| What do you have on hand? | Install one or two of these |
|---|---|
| Nothing at all—I just want to come up with one idea. | anything2explainer (Explanation) / ArcReel (Story) / If you want it all, go with OpenMontage |
| There's video but no audio—I'm relying on the subtitles. | video-shotcraft (real-life footage style) / html-video (data) |
| I recorded this voiceover myself; I need to cut out the filler. | claude-video-editor (Screen Recording Tutorial) / herk-edit-kit (Vertical Vlogs) |
| I recorded it, but I'm too lazy to explain how to edit it. | OpenStoryline (it decides on its own) |
| I want precise control, but I don't want to learn how to use the software. | OpenChatCut |
| Every time I release an album, I have to transcode and resize the files. | ffmpeg-skill + OpenClip |
When listed in order, these five steps form a complete workflow: generate content if there’s no footage, match the rhythm to the footage, edit out unnecessary dialogue if there’s voiceover, and then transcode and trim the video after editing. If you want to use tools to automatically edit AI videos, you could theoretically go through the entire process, but there’s no need to install them all—most people get stuck at one specific step, so installing just the tool for that step is enough.
A couple of final reminders: Most of these skills rely on a local FFmpeg and Remotion environment, so make sure you have all the dependencies set up before installing—otherwise, an error on the very first line will deter half the users. Also, the number of stars changes daily, so take a quick look at the most recent commit date before adding it to your repository—even if a project has a lot of stars, use it with caution if it hasn’t been updated in six months. This is especially true for AI-powered video editing tools; if the editing software’s API changes, older repositories are essentially rendered obsolete.









