From Head of Eng to Indie Hacker. Building AI Story Hub Because My Brain Needed Git & a Filesystem for Storytelling
For years, my professional life as a Head of Software Engineering has revolved around building complex systems. My mental model is structured: think modular components, clear organization, robust version control (thank you, Git!), and reliable deployment pipelines. It's about managing complexity through structure.
But when I'd switch gears to my creative passion – writing stories, building intricate fantasy or sci-fi worlds – I felt like I was stepping back into the digital Stone Age. My writing process felt like a chaotic mess of scattered documents, inconsistent notes, and the terrifying risk of making a change only to realize hours later I'd broken something fundamental without an easy way back. It was the antithesis of everything I knew about managing complex projects.
I needed a tool that didn't exist: an AI-powered platform for creative writing that felt like a well-organized filesystem and offered the non-negotiable safety net of version control, not just for the final output, but for every single component of my story.
The Problem I Had to Solve:
Existing writing software, even good ones, are often built around the linear document model. Great for drafting linearly, bad for managing a growing universe of characters, locations, magical systems, plot points, and stylistic rules that need to remain consistent across hundreds of pages and potentially multiple books. AI writing tools, while powerful, often felt like a black box – unpredictable, and hard to steer consistently for long-form narratives. How do you ensure the AI remembers your obscure world rule or a character's specific speech quirk from 50,000 words ago?
My engineering brain kept asking: Why can't I git commit my character's backstory? Why can't my setting details live in their own structured 'directory'? Why is there no version history for that crucial plot outline I just tweaked?
The Build: Applying Engineering Principles to Creativity
So, I decided to scratch my own itch and build AI Story Hub (https://aistoryhub.co). The core philosophy from day one was simple: bring the power of structured data management and version control from software engineering into creative storytelling, and integrate AI as a controllable, contextual co-pilot.
Story as a Filesystem: The platform is built so that every character, location, item, and importantly, configuration setting (
config.genre,config.tone,config.bannedwords,config.aifor deep control) is its own distinct, manageable entity. It's a structured database disguised as a creative writing environment. This modularity is key to managing complexity.Element-Level Versioning: This was non-negotiable. Every single one of those distinct components – a character profile, a scene, a configuration file – has its own version history. You can make radical changes to a character's personality or rewrite an entire scene, knowing you can instantly revert to any previous version if needed. This eliminates the fear of experimentation.
Controllable AI via Context: The AI isn't just prompting a large model blindly. Our architecture (built with TypeScript, Svelte, Vite, and Firebase) includes an intermediate step where we gather ALL the relevant context from your structured story elements (the character's defined dialogue style, the active config files, the details of the current location, the last few scenes) and use a model like Gemini Flash to optimize a prompt that precisely captures your world's current state and your instructions before sending it to the main generation model. This is how we aim for consistency and control.
Launch & Early Traction (1 Month In):
Launched AI Story Hub about a month ago. It's still very early, but seeing initial usage is incredibly validating:
200+ users have signed up.
They've started 145 stories and are building 23 distinct universes.
They've collectively run over 1200 AI generations.
It seems other creators also feel the need for more structure and control, even if they don't articulate it in terms of filesystems and versioning! The features users are engaging with most are the ones that embody this control: detailed Character Creation, setting per-Character Dialogue Styles, and generating Scenes that respect the established world rules.
Challenges and Lessons Learned:
Translating Engineering to Creativity: One challenge is explaining the value of structure and versioning to users who aren't engineers. The benefits (consistency, safety, iteration speed) are clear to me, but communicating them effectively through UI and onboarding without making it feel overly technical is an ongoing process.
The "Black Box" Perception: Despite building a system for controllable AI, many users come with the expectation (or fear) of a black box. Educating on how configuration files and structured context actually give them control is key.
Finding the Niche: Creative writing is broad. Focusing on users who build detailed worlds for long-form projects helps, but articulating that specific value prop in a crowded AI space is a marketing challenge.
What's Next:
Applying more engineering paradigms! Next up is introducing collaboration features. Because complex projects, including stories, are often best built together. Bringing multiple creators into the same structured, versioned environment feels like the natural evolution.
If you're an indie hacker, especially one who writes creatively or appreciates applying structured thinking to non-code problems, I'd love for you to check out AI Story Hub and share any thoughts or feedback on the approach.
Thanks for reading about the build!
