Holy shit... I just automated sth I thought was impossible with AI: product tutorial videos
The problem at MindPal was pretty simple: we have hundreds of AI templates to share. We know videos of these templates work - some have gotten us tens of thousands of views. But actually making them was a total nightmare.
We tried everything. At one point, we even hired a freelancer, but the feedback loop was exhausting. It actually took longer to give feedback and wait for revisions than it did to just make the video ourselves. It was slow, expensive, and impossible to scale.
When we did it ourselves, it was a massive grind:
→ Record the screen of the behind-the-scene agent builder
→ Record a demo of the agent working
→ Write a script that didn't sound like a robot
→ Record a voiceover or an avatar
→ Spend hours editing everything together
If my co-founder or I were tired or busy, the videos just didn't happen. I assumed this was just the "manual tax" you had to pay for quality.
Last weekend, I got fed up and asked Claude if I could just automate the whole damn thing.
Turns out, I can.
So I spent the weekend cooking something - an internal AI SOP to turn any workflow URL (yes, from just a single URL) into a publish-ready use case video that passes all quality standards in ONE GO.
Here is the new setup:
→ Playwright: Records the screen and even moves the mouse like a human
→ @Claude by Anthropic: Writes the narrative based on our actual product info
→ @HeyGen: Creates the avatar and voiceover
→ @Remotion: Programs the entire edit - syncing everything into a final file
→ @Zernio + @Railway: Automatically publishes the video and saves the assets.
Now, I just give the system a URL and a finished video comes out. I don't even have to click "upload."
I just wrote a post sharing the full behind-the-scenes build, the architecture, and the logic behind of this AI video agent. Check it out here if you think this could be helpful for your company: https://mindpal.space/article/ai-video-agent
P/s: This is what I wake up to every day now 😎✌️
Replies
Does it work better than hiring freelancers now?
MindPal
@everett_thomas Depending on your priorities. If you prioritize the quality of the editing, then this system is definitely not on par with professional video editors. However, for us, we prioritize the feedback loop and the speed of iteration more, and this system provides us with that, helping us shorten the time from the idea to published video.
How scalable is this for enterprise user?
MindPal
@delaney_rose1 what do you mean?
This is incredible. You’ve basically built a 'Video Factory' for your SaaS. 🚀
I’ve been wrestling with the exact same 'Video Tax' while launching my latest Mac tool (Mirowl), and this is the inspiration I needed to stop doing things manually. The Claude + Remotion combo is a game-changer. Thanks for sharing the behind-the-scenes!
MindPal
@safiullah_mohamed I'm glad you find this sharing useful. Try this out and let me know how it goes. ;)
Ovren
This is exactly the kind of AI execution people underestimate, not just generating ideas, but actually removing a painful workflow end-to-end. The “feedback loop was slower than doing it ourselves” part is way too real. Great build.
MindPal
@mikita_aliaksandrovich Thanks, Nikita. The faster feedback loop is definitely one of the best parts of the system. My speed of iteration has gone from months/weeks to days now.
How consistent is the final video quality?
MindPal
@yara_simone Pretty consistent, I think, at least according to my current standards. It's definitely much better and faster compared to hiring an external video editor who doesn't have a lot of context about our products and needs a lot of back-and-forth feedback. However, since this is still a new system, sometimes it also runs into hiccups such as unsuitable visual assets or poorly timed footages but I am currently avoiding that by adding a human in the loop checkpoint where the AI would wait for my review and approval after its initial draft before publishing.
Can this handle longer tutorial videos too?
MindPal
@isla_catherine I think so. I designed it around short-form videos because for us, that's specifically the kind of videos that we want more volume right now, but you can definitely use the same playbook for longer tutorial videos.
Do you still see edge cases where it fails?
MindPal
@ethan_marshall The core flow always works, but it does sometimes run into hiccups like unsuitable visual assets or missing cursor in product footages and so on. It's an easy fix because I can always chat back to the AI to ask it to correct the mistakes.
I’d love to see how it performs compared to your manually created videos.
MindPal
@mathew_chang On an individual level, I think there isn't any video that surpasses our highest-performance video yet. But as a whole system, it is definitely much more consistent, reliable, and allows for much faster iteration compared to before.
Wannabe Stark
MindPal
@jasonhowie Glad to hear that. Have you tried something like this for your company?
Wannabe Stark
MindPal
@jasonhowie What kinds of videos do you often make? I'm happy to share with you what I know, if that helps
This is honestly where things are heading.
You basically rebuilt the entire video production stack into a pipeline — and that matches what we’re seeing more broadly. AI video tools are already cutting production time and cost massively, sometimes up to ~70% compared to traditional workflows
What’s impressive here isn’t just automation — it’s removing tool switching. Script → recording → editing → publishing all in one flow is the real unlock.
Also interesting that you used Playwright for “human-like” interaction. That’s the missing piece in a lot of these setups.
Feels like this is the early version of something much bigger:
content systems, not content creation.
Curious how it handles edge cases though — like UI changes or flows that aren’t clean.
MindPal
@alpertayfurr The system is orchestrated by an LLM, so it can be flexible. As long as I provide it with enough context about what exactly it's doing, it will find its way to accomplishing the task.