I’m building https://TaleTwister.com – an AI tool that generates and narrates bedtime stories based on a kid’s interests, a specific moral lesson, age, and optional specifications. The product was born out of necessity (a stint of bad behavior that comes with age).
It uses GPT-4 for story generation, ElevenLabs for narration, and a simple Next.js + Supabase stack for the app layer. I’m experimenting with story memory (so kids can revisit recurring characters) using vector embeddings, and building a “choose-your-own-adventure” mode with dynamic audio rendering.
Biggest challenge so far: aligning narration, ambient sounds, and story pacing without sounding janky or robotic. Solved it by tokenizing and chunking the story for synchronized audio stitching via ffmpeg.
Another challenge was the inconsistent image illustrations via Dall-E 3. I’ve adopted a dynamic prompting method that includes as many details about the scene, character details, and other visual elements which should remain consistent on each of the storybook pages.
If your kid ever demands “one more story” after a long day, I built this for you. It’s had a meaningful impact on my son’s behavior.
That's really cool! I made something similar but much simpler using R Shiny.
Edit: Looking at the featured story, I notice that the character illustration is not consistent currently. I got relatively consistent character illustration from scene to scene by asking the LLM to write a prompt for image generation based on the story text and use the same character description in each sentence. That got me pretty consistent character drawings. Also, to keep the same drawing style, my prompt was like:
Drawing style e.g., Digital painting {drawing prompt}
Thanks for the feedback! I’ll be incorporating that. The scene is described with the dynamic image prompting however there’s nothing on style in the prompt.
I’ll update the prompt and see how it works. Thanks again!
Echoing other posters, I no longer can shop with confidence on Amazon. The ease of making a purchase is dead and the fast shipping doesn’t counteract the reality that there is no quality control.
Having to sort reviews by most recent and really study the product and seller defeat the point. That’s why I’m back to ordering online with Home Depot and the like.
I agree with you 100% that it's on the individual to find a job which is most suitable for them.
Interestingly, for myself the more intense the day at work is, the more energized I feel when I get home to work on my own passion projects and businesses. Since my role has been relatively autonomous, I'm the one setting the intensity level and if it's a high pressure day it means I bit off more than I bargained for.
To avoid the after-hours productivity slump, I try not to let work get "slow" by taking on new projects, asking other teams what problems they're encountering, and finding new tools to add to my arsenal. That being said, there are still days that seem to crawl by and very little is left to be done. Those are the days that drain me mentally and physically. Some of the people I've worked with have talked about how they spend their workdays pretending to work while actually playing mobile games or whatever and how it makes the day go by faster. I'm genuinely fascinated by what I perceive to be their complacency and lack of ambition.
As background, this is my first role at an organization I'm not a founder of and I started there three years ago at the age of 27. Starting at age 14 I had my website network which was racking up $500+ a month in server bills, so the entrepreneurial lifestyle (rollercoaster ride) is my definition of normal. As a first-time employee, I've had some embarrassing moments while learning the rules and etiquette which is probably common sense to everyone else.
What do you recommend for aggregating and scraping the data? I’ve been working with PyCharm and BeautifulSoup4.
Also, any suggestions for the best ways to apply the data to a website if the data is being refreshed daily? I’ve been using csv files to pass the variables into a Wordpress theme / post but it seems like building something from scratch would be more efficient in the long term.
If you use Chrome, check out the "How your ads are personalized" page. It gives you a glimpse of how accurate Google's targeting is based on all the data they've collected.
I went to a great Google Fireside Chat with Michal Kosinski on data privacy. He said that the only way to get around big data is to spread misinformation. Chose the opposite of your tastes, search for things you don't like, etc. https://www.youtube.com/watch?v=VUwBcTgzbtU
It uses GPT-4 for story generation, ElevenLabs for narration, and a simple Next.js + Supabase stack for the app layer. I’m experimenting with story memory (so kids can revisit recurring characters) using vector embeddings, and building a “choose-your-own-adventure” mode with dynamic audio rendering.
Biggest challenge so far: aligning narration, ambient sounds, and story pacing without sounding janky or robotic. Solved it by tokenizing and chunking the story for synchronized audio stitching via ffmpeg.
Another challenge was the inconsistent image illustrations via Dall-E 3. I’ve adopted a dynamic prompting method that includes as many details about the scene, character details, and other visual elements which should remain consistent on each of the storybook pages.
If your kid ever demands “one more story” after a long day, I built this for you. It’s had a meaningful impact on my son’s behavior.