StoryDrop

Fri, 24 Apr 2026 00:00:00 +0000

StoryDrop

StoryDrop is a mobile-first reading web-app that pairs public-domain books with mood-matched music and light AI enhancements.

You can try it out right now. Click the link or use the QR code.

Scan the QR code to try StoryDrop on your phone

It’s designed to work on mobile first (and it’s still a work-in-progress). Your mileage may vary on tablets and different browsers. I’m busy getting it acceptable on all the major ones.

StoryDrop in action - scrolling through a chapter with interactions.

The Pitch

I pitched this as an idea in early 2025. However, since then everything has been combined with AI - it’s not surprising that a bunch of ‘Book meets AI’ projects have since appeared. That said I think StoryDrop is still quite different from other ideas that have been presented.

Certainly, we can use AI to position the reader inside the story and allow them to talk to the characters and interrogate the storyworld. While I do think that is super-interesting, I wanted to stay close to the original author’s creation.

I just felt that even after all these years ebook readers weren’t really well designed for phones but we have the convenience of carrying our phone with us. While we won’t replicate the great eink screens of dedicated readers I still think there are improvements to be made.

I was fortunate to receive support from CreaTech Frontiers to develop this idea.

The Original Problem - Book vs Noise

The problem I originally set out to solve was that I wanted to read an ebook on the train, in a noisy public environment - so, using noise-cancelling headphones and playing music seemed an obvious choice. However, I found the music choices were not complementary to the book and actually distracting from reading.

I don’t believe this is an issue that affects everyone. I have a similar issue when programming - only certain types of speech-free music is suitable - most I find very distracting and it reduces my productivity.

At the same time I wanted the comfort that we all feel with our endless scrolling (aka doom scrolling). The success of doom scrolling is the skipping things that don’t engage knowing that they will try to hook you in with something else in a few seconds. Skipping sections of a book doesn’t really work - but perhaps there is something about the slicker interface.

So, at the absolute basics: frictionless reading process was the idea - just keep scrolling and complementary music will be chosen.

I’ve done a bit more research recently and have found some apps doing something similar around the complementary music.

Book Content

It would have been great to do this project with a modern book and collaborate with musicians to create complementary music. However, for this prototype my budget is super constrained. So, I used books in the public domain from Project Gutenberg and music from Kevin MacLeod’s Incompetech library.

Public Domain books are slightly tricky as the rules are quite different in the USA and UK. In the USA - it is broadly 95 years after creation for older works. In the UK it is 70 years after the death of the author (other countries use 50 years). So, many that are public domain in the US are not in the UK.

I’m a big Agatha Christie fan and really wanted to use her books in this project. Unfortunately, they don’t enter public domain in the UK until 2047 :( So, watch this space in 20 years time!

Audience Change

The earlier books I experimented with were classic texts: Frankenstein, Shakespeare, Sherlock Holmes etc.

Early experiments - Frankenstein, Sherlock Holmes, & Agatha Christie

For testing purposes it’s easier to have a single book - so I’ve stripped it back and also switched audience slightly.

My own children (currently 11 and 9) - both who are active readers - were very excited helping to playtest our book A Rainbow for Amala - a physical book with QR triggered animations and puzzles. My daughter has read it over and over as she loves the digital interventions.

So, I thought a book from the children’s curriculum would be a good fit. It turns out there is not a recommended reading list on the National Curriculum - but there are various recommendations from teachers.

Here’s a few examples of recommended books also in the public domain.

Alice’s Adventures in Wonderland
Black Beauty
White Fang
Treasure Island

Candidate books from the children's curriculum that are also in the public domain - covers via Project Gutenberg.

So, Treasure Island it is. At the end of this phase I hope to have at least a working web-app that can be used by anyone wanting to read Treasure Island.

Launching StoryDrop

Where can we use AI?

The idea is how AI can be used to enhance the story - but without losing the pure book form.

So, where can we use AI to analyse the text?

Recap the story / chapter
Character descriptions
Provide Dictionary words - lots of older books use old English words / unfamiliar slang
Add chapter questions / reflection
Colorise the dialogue
- Issue with colour choices - do certain colours imply good / bad. How do they change in dark mode?
- Would it help to colourise their name too?

Dictionary popup example

Colour-coded dialogue in dark mode

All of the AI processing is done offline - Claude analyses the source text up front and writes the results out as JSON files that ship alongside the book. Nothing is called at runtime. This keeps the app fast, costs predictable, and means the output can be hand-checked before a reader ever sees it.

What about Game Design?

Narrative is in competition with gameplay - players chase performance, not meaning.

If you’ve looked elsewhere on this site - you’ll see my background is 25+ years in game design of various forms. So, then the natural thought is how can we make the book more game-like. But I’m not a huge fan of gamification in general. And the main issue is not to distract from the original linear book format.

It’s a bit like where I create digital experiences for museums. A mobile phone with access to all the world’s content and connections to the most important people in your life is somewhat distracting when exploring a museum gallery. Games have a similar issue and will distract from the text. It’s a common issue with Escape Room puzzles.

Players become highly focused on solving problems (as that is what we are measuring - performance / speed) to the detriment of engaging with and understanding the story.

So, I wanted to keep the interactions light and not too distracting from the text.

I tried some ideas but for now have settled on classic rewards and collections. I did consider generating this content with AI - but again this isn’t in the spirit of using AI in this project. Instead - I wanted to use AI to analyse and enhance what is already there. So, for now players unlock content - the content that you normally get for free in the original book - the existing illustrations.

I don’t have a good idea for a ‘game-work’ mechanic. In a standard puzzle game - I (as the game designer) can make you play the puzzle again or give you additional challenges in order to ’earn’ the required resources to continue. I’ve yet to figure out a way to make this work with reading that didn’t feel like a deliberate punishment. While players are willing to repeat tasks in order to ‘grind’ towards rewards that they want - I don’t think re-reading the chapter feels fair.

There’s a small amount of collection during the chapter and then an opportunity to complete activities at the end of a chapter. These are still quite light touch - I’d like to explore this more - and perhaps ramp up the game-ness of it to see how it would work.

Collectibles - use coins to buy items.

End-of-chapter activities

Integration with your ebook collection?

My earlier plan was to keep the source ebook material entirely clean - and then provide a set of supporting content (JSON files with chapter analysis, music files and additional graphics). The idea being that in the future this would be an add-on pack for ebooks that you already own. That proved cumbersome for colour-coding the dialogue - so to simplify that I’ve embedded within the original text.

The AI side is handled offline by Claude: I run the source text through a pipeline that produces chapter summaries, character notes, dictionary glossaries and dialogue tagging, all written out to JSON alongside the book. At runtime the web-app just reads those files - no live API calls - which keeps latency low, costs fixed, and lets me review the output before it ever reaches a reader.

Storydrop running on iPad in fullscreen. This is using the focused view where additional text is blurred.

Running on an Android tablet

Next steps

It’s entirely free and ready for feedback. So, please try it, share it and get in contact. I’d love to polish this experience and then add more book requests.

My desire is that it becomes a tool that primary and secondary school teachers can use with their students - but perhaps there are other use cases?

Or, if you are an author with a book you’d like converting to it - get in touch. It would be amazing to work with authors / musicians for a very custom experience.

Thanks

Huge thanks to CreaTech Frontiers for their Foundations grant that made this prototype possible, to Project Gutenberg and Kevin MacLeod for the source material, and to my family for the patient playtesting.

Particular thanks to those who supported me through CreaTech Frontiers: Professor Sylvester Arnab (Professor of Game Science, Coventry University) and Dr Petros Lameras (Associate Professor in Ludic Design, Coventry University) for their valuable insights along the way.

Reading on John Sear

StoryDrop

StoryDrop

The Pitch

The Original Problem - Book vs Noise

Book Content

Audience Change

Where can we use AI?

What about Game Design?

Integration with your ebook collection?

Next steps

Thanks