The knowledge graph links writings, research, projects, and topics so navigation and retrieval understand relationships, not just keywords.
In my daily workflow
- I maintain consistent slugs, tags, and cross-links in content.
- The graph API exposes entities for visualization and sitemap generation.
- Retrieval uses graph hints for related content suggestions.
- I fix orphan nodes when content moves or renames.
How it makes me work smarter
Flat search misses 'how does this automation relate to that skill?' Graph edges answer traversal questions — related gallery topics, breadcrumb context, discovery paths. It complements vector search rather than replacing it.
My setup
- Knowledge graph builder in `features/knowledge-base/lib/graph`
- JSON API with cache headers for public consumption
- Breadcrumb config driven by content relationships
- Sitemap integration for discoverability
On this portfolio
The knowledge graph powers discovery across writings and research and informs SEO structure. Gallery related-topic links mirror the same mental model — explicit edges between AI-first concepts.
