llm-wiki

To build a markdown knowledge base, you need a tool that ingests raw sources like articles and PDFs to automatically create structured, interlinked markdown files for entities and concepts. This approach organizes research workflows into a persistent wiki directory that synthesizes content for querying.

Yes, you can ingest raw sources such as PDFs and transcripts into an interlinked markdown wiki. The ingestion process captures the raw content and automatically generates structured markdown files for entities, concepts, and comparisons to build a persistent knowledge base.

You can check a markdown wiki for broken links and orphan pages by running a lint and health check. This process detects orphan pages, broken links, stale content, and taxonomy violations to ensure the interlinked knowledge base remains clean and consistent.

Yes, you need to configure a specific wiki directory path to create an AI-driven markdown wiki. The knowledge base uses this configured directory to store markdown files, frontmatter, and link structures that organize your indexed entities and concepts.

The best way to query compiled knowledge from a markdown wiki is to use built-in retrieval features that synthesize content from indexed files. This answers questions directly from the persistent knowledge base, reducing the need for repeated manual searches across scattered documents.

Using markdown files for an AI-driven research wiki provides a persistent, interlinked structure that grows over time without database overhead. Markdown supports frontmatter and link structures to organize entities, concepts, and comparisons while remaining easily accessible for health checks.