This blog is dedicated to sharing the coolest ideas we worked on at Considdr (2014–2020) — a search engine that leveraged natural language processing to uncover important insights in full-text documents and provided a tool for users to build, visualize, and continuously update more evidenced-based beliefs based on those insights. “Belief Trees”, “Logical Aggregation”, and “Summarization by Adjacent Document” are among the most novel ideas we developed.