LoreBlendr.AI

• 

  LoreBlendr.AI Home

Inspiration

The inspiration for LoreBlendr.AI came from a simple question: what if AI storytellers could remember you? We love interactive fiction and AI chat, but the experiences often feel ephemeral. Conversations are stateless, and the narrative resets with every session. We were inspired by the rich, persistent worlds of tabletop RPGs and wanted to bring that feeling of a continuous, evolving saga to a mobile experience. We envisioned an AI that wasn't just a text generator, but a true Dungeon Master—one that remembers your choices, learns your character, and weaves your personal history into every new adventure.

What it does

LoreBlendr.AI is a dynamic storytelling engine for iOS that creates deeply personal and endlessly replayable narrative experiences. At its core, the app allows users to:

• Import and Manage Characters: Using a flexible CharacterCard format, users can bring any character to life.
• Build Living Worlds: The Lorebook system lets users define key places, items, and concepts, which the AI actively incorporates into the story. This is compatible with existing community lorebooks.
• Experience Persistent Memory: Our custom MemoryManager ensures the AI remembers key details from your past conversations, creating true narrative continuity and "stickiness."
• Direct the Narrative: Users can define high-level Story Arcs that guide the AI, giving them authorial control over the tone and direction of their adventure without sacrificing the magic of emergent storytelling.

How we built it

LoreBlendr.AI is a native iOS application built entirely in SwiftUI, leveraging the latest frameworks like SwiftData for on-device persistence. Our architecture is designed for modularity and scalability.

The narrative engine connects to any OpenAI-compatible API, allowing users to choose the LLM that best fits their needs. The real magic happens in our ContentViewModel, which orchestrates the complex interplay between user input, chat history, lorebooks, and our persistent memory system to construct a rich, stateful context for every AI request.

For monetization, we integrated the RevenueCat SDK from day one. This was a massive accelerator. Using PurchaseService.swift, we were able to implement and test our subscription offerings in a fraction of the time it would have taken to build the infrastructure ourselves. This allowed us to focus on what we do best: building a next-generation storytelling experience.

Challenges we ran into

Our biggest technical challenge was managing LLM context. Sending the entire chat history for previous sessions with every request is inefficient and expensive. We overcame this by developing a sophisticated MemoryManager. It creates concise summaries of key events and user decisions, which are then injected into the prompt as needed. This provides the AI with long-term memory without creating an unmanageably large context window.

Another challenge was using a CharacterCard specification that was both powerful enough to create nuanced characters and simple enough for users to adopt. The community has settled on such a specification, which allows our customers to import existing content or create their own.

Accomplishments that we're proud of

We are incredibly proud of the feeling of persistence we've achieved. When the AI recalls a detail from a conversation hours or even days ago, it's a magical moment that makes the world feel alive. This "stickiness," powered by our MemoryManager, is our key accomplishment.

We're also proud of the modularity of our system. The ability to easily swap out characters, lorebooks, and even the underlying LLM provider gives users an unparalleled level of control over their experience.

Finally, successfully implementing a paywall and subscription logic using RevenueCat in under a day was a huge win, proving the power of their platform for indie developers.

What we learned

Throughout this process, we learned a tremendous amount about prompt engineering for stateful, long-form conversation. It's an art as much as a science. We also learned the importance of a solid, on-device data architecture; using SwiftData has been a game-changer for managing the complex relationships between characters, lorebooks, and chat sessions. Most importantly, we learned not to reinvent the wheel. Integrating a robust service like RevenueCat for purchases freed up critical development time, letting us focus on our unique value proposition.

What's next for LoreBlendr.AI

We are just getting started. Our vision is to grow LoreBlendr.AI from a personal storyteller into a platform for narrative creation. Our roadmap includes:

• Collaborative Storytelling: Allowing multiple users to participate in the same narrative world.
• Advanced Authoring Tools: Building more powerful UI for creating complex narratives.
• ** Agentic Workflows:** In the future, agents will have a lower latency and can drive the narrative decisions and content while remaining responsive.

Built With

• swift