Knowledge Generator

Custom Knowledge On Demand

Interested in the code? View the repository on GitHub.

The Curiosity Challenge

When my kids ask questions about obscure topics, I face a familiar parental dilemma: Wikipedia doesn't always cover the specific angles they're curious about, yet their thirst for knowledge deserves more than "I don't know." This gap between standardized information and personalized learning isn't just a challenge at home—it's a fundamental limitation in how we access knowledge.

A look at Knowledge Generator in action

What If We Could Generate the Source?

I wondered: could we leverage large language models' vast knowledge to generate comprehensive, structured information about any topic—even highly specific ones? What if, instead of searching for the perfect source, we could create exactly the information we needed, tailored to our specific interests and questions?

Freeform entry to describe what you want to learn

Personalized learning should be as simple as this.

Knowledge That Shapes Itself to Curiosity

Knowledge Generator uses state-of-the-art language models to create detailed, long-form content about virtually any topic, structured as a comprehensive article. Unlike standard search, it doesn't just find information—it synthesizes it around the specific aspects you care about, organizing everything from basic concepts to nuanced details into a coherent whole.

A 10 page report on the history of HCI, tailored to the user's scope of interest.

Learning Without Boundaries

The most fascinating insight? When paired with tools like Google's Notebook LM, this generated content transforms into interactive audio experiences—creating on-demand "podcasts" about even the most obscure topics. For families, educators, and curious minds, this approach fundamentally changes the relationship between questions and answers, making learning feel less like research and more like conversation.

Knowledge generated about geomythology transformed into a podcast using Google's Notebook LM

From One-Night Prototype to Future Vision

Built in a single evening as a proof of concept, Knowledge Generator foreshadowed what major tech companies would later develop as "deep research" tools (the likes of ChatGPT Deep Research, Gemini Deep Research, Grok DeepSearch, and Perplexity Deep Research). While it currently relies on the LLM's internal knowledge (with its inherent limitations), the concept points toward a future where similar tools could integrate real-time search to verify and enhance AI-generated content.

This one-night experiment demonstrates how rapidly innovative concepts can be prototyped and tested, creating functional experiences that anticipate broader industry trends. The real power lies not just in what this specific tool does today, but in how quickly emerging technologies can be shaped into practical solutions that transform how we access, generate, and interact with knowledge.

Let's Make It Real

Together we can transform ideas into testable, functional hypotheses that will give your business a competitive edge.

Request a consultation