This OpenCode Skill Edits Your Videos For Free!

Added: 2026-05-18

[00:00:00]Hey there, welcome to the channel. In this video, we're going to be creating a custom skill to automatically trim and edit videos, which we can then repurpose into YouTube shorts. In order to do this, we're going to be using open code, which is one of the only really freely true open source AI coding agents.

[00:00:17]Currently has around 160,000 stars on GitHub. It is one of my favorite coding platforms, and it's really just as simple as installing it with this one-line command. So, once that's done, you're just going to come over to the terminal. You can see I've just recently started creating YouTube shorts using a custom skill inside open code. So, that's really nice to have. And now, I would also like to do some of the trimming for the videos. Normally, I would use KDEnlive, which is a video editor that's very powerful inside Linux. But in this case, I just want to automate this process using a custom skill, which I can then turn into a YouTube short. I'm just using the web interface, which is really intuitive. It's got a nice interface, and that way I don't have to use the desktop app, which is even easier than inside like cloud code. And code acts don't provide a nice interface like this unless you're running it in the cloud. But in this case, it's just running locally on my machine in the browser, which is pretty cool. It's just creating a new skill under video use.

[00:01:22]And what it does is it transcribes the video using Whisper, detects filler words, finds compelling hooks, and it selects the best 1.5 to 2-minute segment, prioritizes hook in the first 10 seconds, lower filler word ratio, good pacing, question CTA, and all blah blah blah. I think the deal here is that my prompt is not very good. So, I've just gone ahead and created a new prompt, which says, "Create a skill which takes a long-form video and outputs a shorter video no longer than 1.5 to 2.5 minutes long in 16 by 9 format so that it knows that it shouldn't be a YouTube short." And it says, "It should start on a clear hook, remove filler words, and be fine-tuned for virality. And you can see here it's now creating it in the correct directory for open code, so we're going to go ahead and allow that.

[00:02:06]And you can see it has the skill, it's called the viral short cutter. And within a few seconds, it has now created the transcribe.py file, which is a Python file, in order to transcribe the video with word-level timestamps using faster Whisper, and it's going to output the JSON with words, timestamps, and confidence scores. So, if I want to open up this Python file, we can just open that up, and you can see this is what it looks like.

[00:02:32]And basically, it's just transcribing the video and putting it into a file, which it needs in order to add subtitles and figure out where the right hooks are, and also to remove the filler words.

[00:02:44]And by the way, I'm using Quinn 3.6, which is a completely free model inside open code Zen. So, this is costing me zero credits to build this. Yeah, again, I don't actually want a short, I just want the long-form video, and it looks like it is going to output it into 16 by 9 video optimized for virality. That's what I want. And now it's creating that file, and we can take a look at this one. It's very, very simple here. You can see it's only about 100 lines long. It looks for a compelling opening, so this is how it's identifying the hooks.

[00:03:16]And it's also got the filler words in here as well, like um, uh, uh, so it will remove those. And overall, this looks really, really good. So, in this particular the case, I'm going to be running this on my latest video for YouTube. It's about 12 minutes long, and I just want to trim it down to be about a minute and a half to two minutes. And you can see here it's transcribing the video, it's removing the filler words, so it found seven filler words and removed them. It's finding a compelling hook, it's selecting the best segments, and it's going to extract those segments, and it should be able to create the video. And hopefully, it's in the right format, and then I can just take that, and then I can use one of my other skills to create a YouTube short. So, you can see it created the viral short cutter skill.

[00:04:01]It tested it successfully, so it transcribed it a test video. It removed seven filler words, auto detected a hook, selected five high-value segments, and it outputted in the correct format 1920 by 1080, a 64-second video, which is 6 MB in size. So, if I come over here and open that up, you can see exactly what it looks like over here, and this is trimmed down to be only 1 minute long, and I'm just going to say "Trim this video."

[00:04:30]And we're just going to paste in the direct path.

[00:04:33]So, now it's going to trim this video, and it should just put it in my LinkedIn directory, which is exactly where I want it to be because I also post these videos on LinkedIn.

[00:04:42]And now it's asking, "How do you want to trim this 10-minute video?"

[00:04:46]So, I'm just going to tell it to create the viral short in 16 by 9 format.

[00:04:50]Should be longer than about 2 and 1/2 minutes long. It's going to use the viral short cutter, which is the skill that we've just created.

[00:04:57]And now it's running that on my open code tutorial, which I've just pulled in over there. After about 5 minutes and a lot of processing on the CPU, it was able to transcribe the video, remove the filler words, find a compelling hook, select the best segments, and output the finished file as the open code video skill tutorial short. So, if I come over to videos here, we should see there it is. Here is the finished trimmed video.

[00:05:27]It's a It's a minute and a half.

[00:05:30]And now what I can do is I can just repurpose this into a YouTube short.

[00:05:34]And ultimately, that's just going to make the job easier for the custom skill so that it doesn't have to go through a full 10-minute long video. We can go ahead and now convert that into an actual YouTube short that is properly formatted. It's got the talking head in the bottom half and then also the screen share portion in the upper half.

[00:05:54]And now, as you can see, it's finding the short video file. It's creating a PIP in the viral short, so it's going to identify where the talking head is. As you can see, it detected where the PIP is located.

[00:06:06]And I've already made a tutorial on the channel showing how to do this. So, if you want to check that one out as well, feel free to do so.

[00:06:13]And now, it's converting the viral short to a 9 by 16 YouTube short, and it's going to put it in the same directory over here. So, as you can see, it's already creating the YouTube short within a few seconds, completely for free, again, using open code.

[00:06:27]And it's done. So, that took about 30 seconds, probably slightly less than 30 seconds. It says done, the video has been converted to a YouTube short. Says exactly where it's located. The PIP was detected, the screen content plus the talking head are stacked vertically. So, now if I come over here, I can preview it, and sure enough, there we go. And the talking head's in the bottom half, and the screen share portion is in the upper half. This is well-formatted for a YouTube short. And now also, if I wanted to add subtitles to this, I also have the option to do that as well.

[00:06:59]So, I could basically just take the transcript that it's already gotten using Whisper, and I can tell it to add subtitles to the YouTube short. Add subtitles to the YouTube short in the middle of the video with animations, and it can just transcribe the video and then add the subtitles in. I want them to be displayed right in the middle here between the screen share and the talking head.

[00:07:21]Should be able to do that quickly and easily over here. And then I could also build that directly into my custom skill as well if I really wanted to. And overall, I'm really, really happy with Quinn 3.6. This is honestly one of the fast fastest free models that I found so far.

[00:07:39]Even better than MiniMax 2.5 in a lot of ways, which I was really happy with previously. That one's also completely free. So, highly recommend using that one inside Open Code.

[00:07:50]And now it's creating a a separate Python file to add subtitles to the video under the Viral Short Cutter. And basically what this is doing is just going to add the subtitles onto a video, and it's built directly into my Viral Short Cutter. So, if I want to add subtitles, I can also go ahead and tell it to add the subtitles to the YouTube Short, which is exactly what it's doing right now. You can see it's adding the animated subtitles to the YouTube Short.

[00:08:16]So, it's just going to transcribe the YouTube Short. It's going to create the animated subtitles quickly and easily within a few seconds, completely for free inside Open Code.

[00:08:25]How cool is that? And yep, sure enough, it now has the subtitles in the middle of the screen here. It does need a little bit of work, but overall you can see it is able to add the subtitles and burn them directly into the video. I would want these subtitles to be a little bit bigger. I want the animation to look a little better as well. But overall, still work in progress, but you can see what we're able to do with these custom skills inside Open Code, which is available completely for free. You don't have to spend any money on credits like in Cloud Code and Codex. You can just use Open Code to do all of this and just use one of the many available free models. I have a lot of different providers connected, so as you can see, literally got hundreds of models available to choose from in here.

[00:09:12]And with that being said, thank you so much for watching this video. Please go ahead and subscribe to the channel, and also go ahead and check out my free school community over here. We've currently got 2,700 members in here.

[00:09:24]It's growing fast. You can join and check out the classroom where we have lots of tutorials and courses and resources available for you, as well as go ahead and introduce yourself inside the community tab. We've got new people introduce themselves every single day.

[00:09:39]So, go ahead, check out the school community, and with that being said, thank you so much for watching this video. Please go ahead subscribe to the channel if you haven't already. Take care, stay safe, peace from Thailand, signing off.