The 7 Levels of AI Video Editing: Where Are You Really?

Some of you reading this feel like AI is completely overwhelming. There’s a new tool every single week, it’s moving faster than any of us can keep up with, and frankly you feel frazzled. Others of you feel pretty confident. You’ve got a grip on AI, but you’re wondering how deep this goes and whether there are workflows you’re missing. The good news: this is for both of you. These are the seven levels of using AI in video editing, so you can see exactly where you’re at and where you might go next. There’s no judgment here, only progress. And as we go, you’ll notice a theme: if you want to stop renting your intelligence, there’s a way to do that. Stick around for level seven.

The short version: Most editors sit at level two or three, using the AI features built into their editor and dabbling with a plugin or two. The real shift happens at levels six and seven, when you move from subscribing to other people’s tools to building and owning your own pipeline with open-source tools and local AI. That jump is far easier than you think, and you don’t need to write a single line of code.

Why I got hooked on AI editing in the first place

The first time I used AI in video editing was one of those rough web apps that took a long-form YouTube video, cut it into clips and slapped branded captions on, tools like Opus Clip. I remember thinking it was insane that something could do all of that so quickly. Later, I found tools that could handle all the fiddly metadata for my stock footage uploads, and I calculated a single upload saved me about four months of my time.

My goal, and the goal of this whole channel, is to build a job where most of the difficult, boring stuff is gone, and what remains is fun, creative work with a truckload of passive income from assets we’ve already made. That’s why I like AI: it can actually take me there. So, the levels.

Level 1: The AI avoider

You don’t just avoid AI. You hate it. You feel like using it damages your integrity and the quality of your work. Cutting silence by hand, doing your own rotoscopes manually, fills you with a sense of pride that this was bespoke, handmade. And look, I’m not here to call you out. But the sad truth is that nobody will notice you cut the silence by hand on that long interview edit.

If this is you, the fact you’re reading this is probably the first step. You can’t wind back the clock, but you can acknowledge that there’s a bright future for you in video editing, and you’re probably not as far away from it as you think.

Level 2: Built-in AI, reluctantly

You use the AI features inside the apps you already own: CapCut’s auto captions, Final Cut’s Magic Mask, DaVinci’s incredible AI slow-mo. Maybe you use ChatGPT to organise your thoughts or draft a script, but it goes no deeper. Honestly, this is where most people are. Maybe you’re put off by AI, maybe a little scared of it, or maybe, if we’re honest, you look down your nose at people who use it.

If you can put your ego to one side, your life gets a lot easier. These tools save you time and make you more money, and getting from here to levels three, four and five is much easier than you’d think.

Level 3: Curious but cautious

You’ve started trying plugins and web apps that have genuinely changed how you approach editing. You’ve seen the Base44 ads, your friends talk about Replit or Claude, you’ve downloaded Gling or Descript and given them a go. You’re willing to try, but hesitant. You’ve heard the horror stories, you know there are trade-offs. And yet there’s that nagging feeling you might be missing out.

One day you catch yourself thinking, “I won’t use AI because I’ve always done it this way”, and then you realise that thought process isn’t the most sensible. You were in your element in 2016 watching Peter McKinnon teach transitions, and you’re a tiny bit scared of this brave new world. That’s okay. The fact that you’re curious enough to wonder how deep the rabbit hole goes is a very good sign.

Level 4: AI becomes daily use

This is where AI moves into your daily workflow: audio, story, video and colour all at once:

• Audio: ElevenLabs for the voiceover line you forgot to record or for dubbing into another language; Adobe Podcast to rescue a windy location take (which, by the way, is insane).
• Reframing and cleanup: Topaz for upscaling, auto-reframe to turn 16:9 hero edits into vertical clips without touching a keyframe.
• Pixels: DaVinci’s Magic Mask and Runway’s rotoscoping. No green screen, no keyframing every second by hand. I remember rotoscoping inducing genuine rage. Now you press a few buttons and review what the AI made.
• Long-form to clips: Opus Clip cutting up your interviews.

You’re dramatically faster at turning around edits, and you’re doing things that didn’t seem possible a year ago. But you’re starting to realise something important: this is more than a tool. It’s a skill, and you’re still scratching the surface.

Level 5: Generate everything

Level five takes a different route, and honestly it won’t apply to everyone. This is generative AI: Sora, Veo, Kling, Runway and Luma creating footage no camera ever saw. It’s easy to generate AI video with a couple of prompts. It’s really hard to do it well, which is why this sits at level five. If you can blend generated shots into real footage convincingly, a world of possibilities opens: the drone shot you didn’t have permission to get, the audio you missed on the day, custom music from Suno instead of another royalty-free site, your own voice cloned in ElevenLabs.

This is the one-person-Pixar moment: you realise you could make a full trailer-style piece in your bedroom for as little as $50. And you start seeing ads for tools like Vía or Eddie AI promising prompt-style editing: matched B-roll, multicam cuts, colour grading, the lot. Maybe you’ve even subscribed to one. But here’s the question that leads to the next level: what if you didn’t need to subscribe? What if you could build something like this yourself?

Level 6: Open source, the biggest shift

Level six, to me, is the biggest shift of all. It’s the moment you realise that all this time you’ve been using other people’s tools, and it’s been costing you a huge amount of money. Then you discover the open-source world, projects like HyperFrames, and it doesn’t just change what you can do, it changes your entire perspective: video editing is just code, and these tools work with code. They’re free. Nobody is holding a subscription over you.

Maybe you already use Claude as a chatbot but haven’t touched Claude Code because you think, “I don’t know anything about coding.” Here’s the thing you haven’t realised yet: you don’t need to know a single line of code to use it. At this level you stop being a passive customer and start being an explorer. There are steep learning curves, but the payoff is unbelievable. It will feel like you’re doing a different job to everyone else in your industry. The people at level six are the ones who are going to win the AI race.

Level 7: Owning your pipeline

Here’s where the whole thread pays off. Almost everything from level five lives behind a credit meter or someone else’s server. Level seven is when you’re not renting anymore: this is your intelligence, and you own the rig that runs your editing.

It starts with a curious moment: you ask Claude or Base44 whether it’s possible to make an app that edits X, Y and Z, and to your amazement it says, not only possible, but doable. So you start building your own AI software into your workflow. Your own plugins. Your own effects. Beautiful motion graphics with tools like Remotion. And then maybe you discover local models like Ollama and realise you’ve been paying other people to use the RAM on your own computer to generate intelligence. If you’re anything like me, that’s a slap-yourself-on-the-forehead moment.

Now you plug local AI into your own apps and power your entire system for free. Every niche problem your specific clients throw at you is one app away from being solved. You literally ask Claude Code, “Can you make me this application?” What used to be a bloated subscription stack becomes a lean, efficient weapon. Every ad for a shiny new AI tool becomes exciting for a different reason: you know you could probably build it yourself. There are no gatekeepers anymore.

And here’s the real prize: when you sit down to edit, it’s only to edit the things you enjoy. The boring tasks are automated, deadlines stop crushing you, and you get breathing space to focus on the only things that actually separate you from everyone else: taste and story.

This is exactly why I build my own tools. My AI editor Edit Pup learns how I cut and drafts in my style, runs on my own Mac, and my footage never leaves my machine. If you want to try the apps I’ve built, or you want the stories and workflows from this journey each week, join the waitlist.

Is there a level 8?

There’s probably a level eight, a level nine, maybe a level ten. Truthfully, I’m probably not even fully at level seven myself. It’s something I’m working towards. The great thing is we’re on this journey together. If you think I’ve missed a level or a key tool, leave a comment on the video. It’s a small channel, I read and reply to everything.

And if level seven has you curious about Remotion, making custom motion graphics from plain-English instructions, for free, on your own computer, that’s exactly what the next video and article cover. No technical chat, because I’ve never written a line of code in my life. I don’t vibe code; I survive code.

Frequently asked questions

What are the 7 levels of AI video editing?

1) The AI avoider, 2) built-in AI used reluctantly, 3) curious but cautious with plugins, 4) AI in daily use across audio, video and colour, 5) generating footage, music and voices, 6) discovering open-source tools, and 7) owning your entire pipeline with your own apps and local AI models.

Do I need to know how to code to reach the top levels?

No. Tools like Claude Code let you describe what you want in plain English and build working apps and plugins for you. I’ve never written a line of code in my life and I build my own editing tools.

Isn’t AI-generated video (level 5) risky for client work?

It can be. Generated content increasingly carries labelling and disclosure obligations, especially for EU audiences. The higher levels here are mostly about automating your workflow, not generating your footage, which avoids that problem entirely.

What’s the difference between level 6 and level 7?

Level six is discovering that free, open-source tools can replace your subscriptions. Level seven is actually building your own pipeline on top of them: your own apps, powered by local AI models on your own machine, so you own the software instead of renting it.