Community Discussion · Use Cases

While the kids are asleep, I used WorkBuddy to streamline my AI video side hustle

Xiaoyu's MomXiaoyu's MomAug 312026/08/31 189 views

Hey sisters, I just read an article about AI video generation in 2026, saying that turning text or images into video clips is becoming more like a daily utility tool. The article gave an example where someone used "wild rabbits in backyard surveillance footage" to create short videos, reportedly gaining millions of views within days. My first reaction wasn't wanting to become a director, but rather taking advantage of my baby's nap time to open WorkBuddy and see if it could help me manage my pile of topics, scripts, and assets.

Problem: Generating videos isn't hard; finding inspiration while babysitting is

I've recently started trying Xiaohongshu (RED) and have used tools like Doubao, Jimeng AI, and Kling AI. Honestly, generating a 5-second video isn't as mystical as imagined. What's difficult is: what to shoot today, how to write the script, how to break down storyboards, how to tweak cover titles, which content fits Xiaohongshu vs. Instagram, and how to interpret data after posting.

As a mom, the only time I have is those two hours during the baby's nap and evenings after putting them to sleep. I can't absorb or remember overly complex workflows. So I decided to treat WorkBuddy as a backend assistant: it doesn't film videos for me, but it helps me organize, split tasks, and generate checklists.

Solution: Open WorkBuddy and create an "AI Video Side Hustle" project

My setup is the desktop version of WorkBuddy, though the web version works too.

1. Open WorkBuddy, click "New Project" in the top left, name it AI Video Side Hustle, and confirm. You'll enter the project homepage, where you can see this new project on the left sidebar.

2. Click "New Folder" and create four folders: 01 Topic Library, 02 Scripts, 03 Assets, and 04 Publishing Log. Dump all future files into these.

3. Drag viral screenshots, photos you took, baby food records, and growth milestone screenshots into 03 Assets. I dumped 20-30 files at once; after waiting a minute or two, it categorized them by content summary.

4. Go to "Template Center," select "Document/Table Hybrid Template," and create a new Topic Script Table. You'll see an empty table with just five fields: Topic, One-line Hook, Script Structure, Storyboard, and Publishing Platform.

5. Select the Topic Script Table, click "AI Process," and input a fixed prompt:

Based on the assets, generate 10 AI short video topics suitable for moms' side hustles. For each topic, provide a one-line hook, a 30-second script, 3 storyboard frames, and specify whether it's better suited for Xiaohongshu or Instagram.

A storyboard breaks the video into several scenes; prompts are instructions for the AI; hooks are the sentences that grab attention in the first 3 seconds. After running, the table will populate with about 10 rows of content, which you can edit row by row.

Configuration and Pitfalls: I finally stuck to just three rules

Initially, I made mistakes by trying to get WorkBuddy to do everything at once—topics, scripts, tables, publishing plans. The output looked like template essays with hollow titles. Eventually, I learned my lesson: make it do only one thing at a time.

My current configuration looks roughly like this:

  • Generating Topics: First, ask it to find "real moments" from 03 Assets, such as the baby's first taste of solid food, a two-year-old's catchphrase, or the three-minute meltdown during bedtime routines. It organizes these into a table, and I manually delete the fake-sounding ones.
  • Generating Scripts: Send one specific topic to it individually, requesting "colloquial tone, suitable for filming while watching the kid, narration under 60 characters."
  • Publishing Checklist: Ask it to split into two versions per platform: Xiaohongshu needs a lifestyle feel, Instagram needs visual appeal. But don't let it auto-publish. After splitting, I still manually tweak titles and cover copy; otherwise, it screams "template."

Previously, I found multi-platform publishing tedious. It's still tedious, but WorkBuddy breaks the hassle into checklists, so at least I don't have to memorize things while holding the baby.

I also set permissions. A friend wanted to help me edit covers. I set 04 Publishing Log to "Edit by Owner Only" and 03 Assets to "Viewable by Specified People." She can see the assets but can't mess with the publishing plan.

Every Sunday evening, after the kid is asleep, I do a cleanup: mark published topics as "Used," move unused assets to Old Assets, and copy high-performing scripts to Viral Templates. This small action helps it reference my preferences more accurately next time.

Results: It didn't make me rich, but it reduced my anxiety

Yesterday noon, I spent about 40 minutes picking 6 topics from 18 assets. WorkBuddy generated scripts and storyboards, which I then fed into Jimeng AI and Kling AI to produce visuals. In the evening, after putting the baby to sleep, I only needed to pick one to publish.

This approach is similar to how I used WorkBuddy before for organizing photo albums and making PPTs. Tools improve efficiency, but deciding which moments are worth keeping is still up to me.

Looking ahead, I think over the next six months, the AI video side hustle market will clearly split into two layers: the upper layer competes on model generation capabilities, while the lower layer competes on organization and distribution via tools like WorkBuddy. Ordinary people shouldn't just obsess over prompts; getting your assets and workflows running smoothly is key to sustainability.

2 replies

?
Ctrl + Enter to reply
Shua Ti Zhong

This pipeline looks nice, but how do you prove in an interview that it's not just calling APIs? Algorithm roles seem to value understanding the underlying tech more.

Mo Mo
Mo MoAug 31

As a fellow parent running a side hustle, I totally get the pain of fragmented time. While WorkBuddy is okay for managing assets, if you want to batch produce videos, you still need to rely on the underlying capabilities of the model. Recently, I've been using Wan3.0 for video generation, and I feel that converting documents to video has more potential than just stitching together clips based on duration. Want to try feeding it a script directly to generate storyboards?