Moire
Moire is a personal reflection system that transforms unstructured journal entries into a living semantic map of thoughts, emotions, habits, and identity. Instead of encouraging users to reread old entries, the system surfaces patterns over time — what themes recur, how ideas connect, and how the user's inner landscape evolves.
Journal entries are processed through a multi-stage NLP pipeline that segments text into coherent thematic chunks, extracts high-level topics and sentiment, canonicalises those topics into persistent conceptual nodes, and builds weighted connections between them across time. The result is a dynamic graph where nodes represent meaningful concepts in the user's life and edges represent how those concepts co-occur and evolve together.
The project is driven by a core question: How can reflection support growth without forcing people to manually look backward? Rather than presenting raw logs, Moire acts as a reflective mirror — revealing identity, motivation, struggles, and growth as emergent structure rather than explicit memory recall.
Tools & Technologies
Project Overview
Key Takeaways
- •Designed and implemented an end-to-end semantic processing pipeline using LLMs, embeddings, and graph-based representations
- •Built a canonicalisation system that maps noisy, LLM-generated topics into stable conceptual nodes using embedding similarity, centroid updates, and LLM fallback checks
- •Implemented semantic chunking to preserve context while enabling fine-grained insight extraction without over-fragmentation
- •Used vector embeddings and cosine similarity to manage concept identity, merging, and drift over time
- •Architected a clean separation between API orchestration and core domain logic, improving extensibility and testability
Challenges Overcome
- •Preventing topic explosion and semantic drift from LLM outputs by introducing TopicAlias mapping and node centroids
- •Balancing automation vs interpretability, deciding when to trust embedding similarity and when to defer to LLM reasoning
- •Preserving contextual meaning while chunking entries, especially when themes reappear non-linearly
- •Designing a graph system that evolves incrementally without regenerating or invalidating existing structure
- •Avoiding 'therapy-bot' failure modes by constraining tone, scope, and assumptions in reflective summaries