THE MIRROR

The Problem: We produce thoughts faster than we can understand them.

Knowledge workers, researchers, students, writers, and anyone who journals or takes notes face a universal frustration — they accumulate thousands of thoughts, ideas, and notes over months and years, but have no way to see the shape of their own thinking.

Specifically:

You can't see patterns in your own ideas. You write the same insight three different ways in three different months without realising it's the same thought. You forget what you once cared about. Ideas you obsessed over quietly disappear. You don't notice until years later. You can't find hidden connections. Two concepts you've been thinking about in separate domains might be deeply related — but they live in different notebooks, different apps, different months. Your thinking is invisible to you. The people best positioned to understand your intellectual landscape — you — are also the least able to see it objectively. Who experiences this?

Researchers juggling dozens of half-formed ideas across papers and notes Founders and builders with scattered product thinking across Notion, Obsidian, journals Students who study broadly but can't see how their knowledge connects Writers and creatives whose journals are rich but impenetrable Anyone who has ever thought: "I know I've thought about this before — but where did that thought go?" The problem is not a lack of information. It's a lack of reflection infrastructure — tools that help you understand not just what you wrote, but what it means about how you think.

THE MIRROR is an AI-powered interactive visualization that transforms a user's unstructured thoughts — notes, journals, ideas, meeting notes, anything — into a living, explorable map of their intellectual world. The tagline: "Your thoughts, reflected back to you."

How it works:

• You provide raw text. No structure required — paste your Obsidian notes, a year of journal entries, meeting notes, or stream-of-consciousness writing.

• Claude (Sonnet) extracts the intellectual graph. The AI identifies concepts, clusters them into thematic groups, maps relationships between them, and detects:

Recurring thoughts (what you keep returning to) Emerging ideas (what's growing in importance) Abandoned ideas (what you once cared about and forgot) Hidden cross-domain connections (non-obvious bridges between distant concepts) Temporal evolution (how ideas transform over time) Critically, Claude outputs structured JSON only — the model never touches the UI. The frontend renders this structured representation as a cinematic galaxy visualization.

• You explore an interactive constellation of your thinking. Concepts appear as glowing nodes. Relationships are animated bezier curves. Clusters glow as distinct thematic regions. The visualization feels like a planetarium for your thoughts — beautiful, alive, and explorable.

• You interact with your intellectual landscape:

Click any concept → see its mention count, timeline (first to last appearance), AI interpretation, and source quotes from your own writing "Why are these connected?" → Claude explains the specific relationship using evidence from your text, distinguishing source evidence from AI interpretation Hidden Connections → reveals non-obvious bridges (e.g., "Your meditation practice and your debugging approach share the same cognitive structure: observe without immediately reacting") Forgotten Ideas → highlights concepts that disappeared from your writing, with their last known date Ask the Mirror → ask natural language questions like "What have I been thinking about most?", "What idea keeps returning?", "What contradictions exist in my thinking?" — answered with evidence from your own words How it solves the problem: THE MIRROR makes the invisible visible. Instead of thousands of scattered notes, you get a single explorable map — one that shows you not just what you wrote, but who you are as a thinker. Every insight is grounded in your own words. The AI doesn't tell you who you are; it shows you what you've already said, arranged in a way you've never been able to see before.

The result is the experience the product is designed around: "I gave an AI a pile of messy thoughts — and it showed me something I couldn't see myself."

Claude (Sonnet) extracts the intellectual graph. Critically, Claude outputs structured JSON only — the model never touches the UI. "Why are these connected?" → Claude explains the specific relationship using evidence from your text, distinguishing source evidence from AI interpretation