tl;dr: Yeah, likely not gonna happen.
With no jobs on the app development horizon, and after recently completing a fully functional social site for sharing thoughts and stories, I decided to try my hand at a little short-story animation to tie in with the StoryCore34 website I built. The site’s purpose is to share stories and thoughts and comments in a positive, encouraging environment.
The post below describes the two content creation tools I tried, as well as the StoryStarter feature that was recently added to StoryCore34.
Bottom line: The token usage spent fixing the ‘slop’ largely offsets your best efforts to produce a cost-effective video with minimal fixes.
My initial research brought two content creation tools to the top:
- Magiclight.AI
- The resulting video linked below was created using: Magiclight.AI. As the creator of this short story video, although I enjoyed the tool overall, the cleanup costs in tokens was way too much for my budget, so this will be the only video I will be making with this creation tool. A lot of regens resulted in the same issues, despite massive wording specificity. But mostly, everything costs: while I was running 20-token re-gens and 150-token re-posing trying to get each storyboard right (or worse, having to apply mass changes across all storyboards), It churned through over 7,000 tokens and I ran out when I only had one regen left to run: I had to buy a full batch to finish that one regen.
- Granted, it might could all be worth it if you know how to choose better models or some other tricks of trade, and I’m sure the tool will improve dramatically over time.
- Resulting video: StoryStarter: Oh My Diary — Where you can help decide the story.
- FYI: “StoryStarters” tie into ‘Feature 2’ of the StoryCore34 website, which was completed Sep 5, 2026. StoryStarters are community-driven, paragraph level, ad-libbed stories where anyone can participate.
- Google Flow
- The resulting video linked below was created using: Google Flow + Storyboard Studio. Although Google Flow can introduce a variety of nuanced AI glitches, creating my first video wasn’t too much of a nightmare. The model generator is gonna mess stuff up, so be prepared to pay to fix it, but depending on your level of detail, it usually only took one to three rounds of follow-ups. More detailed fixes can likely take a half dozen or more iterations. And it provides comps up front, which (usually) tend to influence the resulting animation, but paired with an AI agent that can better converse in model-talk on your behalf.
- You may also trip their filters occasionally, depending on your storyline, but you’ll likely find ways to get it to work (given a clean storyline, that is).
- This was the first of a teen sibling mystery series that will tie into Feature 3, “Sharing Literary Content” (stories, poems, etc.), which I’m currently building. This is the first animation of the series, I have the 2nd mystery’s animation near done (waiting on token replenishments), and am partially into writing the 3rd mystery (and learning a tad bit about cotton gins in the process).
- Resulting video: Unpart the Start | Chappter Mysteries — Reversing life’s most unreal mysteries | Nathan and Penny Chappter recount years of unreal mysteries solved in their town, as told by them decades hence.
I had hoped to be able to get my second ‘Chappter Mystery’ story completed in Google Flow, but ran out of tokens. My subscription renews in 3+ weeks, and at 50 gifted tokens a day, it will take another few days to finish. And so the third story in the series will be held off for a bit as well.
For StoryCore34: I just kicked off building Feature 3: Sharing Literary Content, for sharing stories and poems and the like. You can get site updates on when the site gets updated, such as when Feature 3 goes live.
Cheers!
Keith DC | https://keithdc.com
