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
The real win here isn't automation alone, it's removing the feedback loop bottleneck that usually kills consistency.
MindPal
@leah_josephine agreed 👍 that's my ultimate goal
@sylviangth 🙌
If the first draft misses context, how easy is it to correct without rerunning the entire pipeline?
MindPal
@jack_sullivan5 It's as easy as texting a prompt!
@sylviangth Got it! But does that tweak propagate back through the pipeline, or are you just fixing the final output?
MindPal
@jack_sullivan5 We can always ask the system to undo or revert to a previous step, but most of the time the system nails the core part of the video, so most suggestions I have, if any, would be about improving the final output more than changing the direction drastically.
Feels like you intentionally traded peak quality for speed and control, which makes sense for this use case.
MindPal
@wyatt_cameron would you have done the same?
@sylviangth Yeah, In this setup I probably would've made the same trade, but only if I could still plug in a quick manual override for edge cases.
How often does the system break when recording flows with Playwright, especially on dynamic pages?
MindPal
@oliver_nathan2 For us, not very frequent so far. I think the only hiccup I have seen is that the AI forgets sometimes to add a human-like cursor into the final screen recordings. Besides that, it's been doing pretty great. Partly because, while the AI agent template to demo is different each time with different inputs and structure, the overall flow is quite predictable. I still can provide it with a clear goal so it can follow and course correct if needed.
@sylviangth Feels like that predictability is doing most of the work here. Curious how it holds up when flows get less structured.
What happens if the URL has conditional flows or requires login steps. Does the system still handle it smoothly?
MindPal
@paige_lauren1 The URL I provided is from my product, so it does have a login step. I just provided the agent with a username and a password from a developer account to use.
@sylviangth That works for auth, wondering how it behaves once flows start branching or changing mid run.
Does it ever break or need manual fixing?
MindPal
@isaac_dominic1 For now, yes, since this system is only just ten-day old, it is without hiccups, but I added a human-in-the-loop step where it will generate the draft video first for me to review before publishing it to social media channels. This way, if I spot any weird footage or inappropriate visual assets, I will ask it to correct course before proceeding.
Can it handle complex workflows or just simple ones?
MindPal
@fiona_margaret How complex are you thinking of? Right now, this system is automating four types of short-form videos that we make:
AI agent template promotion videos
Turning our long-form user interviews into short clips
Product angles videos
News and trends updates videos
I'm also using this system to edit long-form videos we make, too. But for long-form videos, right now we are still recording the raw footages ourselves; the system just handles the editing part.
@fiona_margaret this is pretty complex as is I think. It just seems simple because its been put across post execution
MindPal
Damn i just lost my job:)
MindPal
@dat_vo_dinh And now you have the space to do better jobs 😁👍
How much cost savings did you actually see?
MindPal
@freya_madison Cost saving is not my main goal for this AI video agent, so I don't really pay attention to it, but two things definitely stand out for me:
I don't have to pay for an external freelance video editor for this, so that's a big saving.
I noticed that the video agent feature on Heygen is actually quite expensive, but for this system, I don't need their video agent feature. I just need their video generation feature, and that's much more affordable.
Can this work for marketing agencies too?
MindPal
@rowan_elizabeth1 Depending on the specific kinds of videos you make, I definitely think some parts of this system can help with your existing video making workflows. If you can share more about the videos you often make, I can help share some tips and insights based on my experience.