What problem does it solve?
Maintaining a consistent knowledge graph of search and information retrieval research requires manually creating notes with correct frontmatter, cross-links, and grounded claims, which is error-prone and inconsistent across entity types.
Core Features & Use Cases
- Entity extraction and normalization: Identifies Concepts, Topics, People, Companies, Tools, Conferences, Case Studies, and Datasets from source text or existing vault notes, with canonical naming rules to prevent duplicates.
- Structured note authoring: Creates or updates notes in the correct folders with type-specific frontmatter fields, Obsidian wikilinks, and Related Notes sections.
- Mandatory grounding checks: Verifies every claim, date, and relationship against the source material before saving, with no grounding markers left in the note.
- Use Case: After processing an article about SPLADE, extract the concept, its authors, and the tools mentioned, then create linked notes for each entity and hand off to the graph maintenance pass.
Quick Start
Write a topic note for learned sparse retrieval pulling together what we already have on SPLADE and related concepts in the vault.
