reading-assistant

Generate Obsidian Zettelkasten notes from existing outputs

Extract facts, examples, metaphors, quotes, and glossary

Build or rebuild sqlite-vec RAG database

Set up reading-assistant: create venv, install dependencies, validate environment

Full pipeline: parse, summarize, extract, review, QC, Obsidian notes, RAG database

Search the RAG database

Fetch and synthesize book reviews

Load and validate reading-assistant configuration

QC retry orchestration logic

Show pipeline processing status for a book

Generate chapter summaries in source + target language

Generate atomic Obsidian Zettelkasten notes from processed book data. <example> Context: All outputs are QC-verified assistant: "I'll use the cardmaker agent to generate Obsidian notes." </example>

Split chapter text into overlapping chunks for AI processing. <example> Context: Parser has produced chapter files assistant: "I'll use the chunker agent to create processable chunks." </example>

Extract structured knowledge from book chapters: facts, examples, metaphors, quotes, glossary. <example> Context: User wants to extract key insights user: "/read:extract ~/books/deep-work.epub --category quotes" assistant: "I'll use the extractor agent to find memorable quotes." </example>

Build sqlite-vec RAG databases for book content search. <example> Context: Book has been processed assistant: "I'll use the indexer agent to build the RAG database." </example>

Parse EPUB files into structured chapter data, metadata, and detect language. <example> Context: User wants to process a new book user: "/read:process ~/books/thinking-fast-and-slow.epub" assistant: "I'll use the parser agent to extract chapters and metadata." </example>

Spot-check extracted items for correct categorization. <example> Context: QC coordinator dispatches category audit after extraction assistant: "I'll use the category auditor to spot-check extraction categorization accuracy." </example>

Orchestrate quality control with parallel sub-agents, auto-retry, and threshold gating. <example> Context: Summaries and extractions are complete assistant: "I'll use the QC coordinator to verify quality." </example>

Verify that chapter summaries cover major topics from source text. <example> Context: QC coordinator dispatches coverage check after summarization assistant: "I'll use the coverage checker to verify summaries address all key topics." </example>

Mechanically verify extracted quotes against source text using fuzzy matching. <example> Context: QC coordinator dispatches quote verification after extraction assistant: "I'll use the quote verifier to check all extracted quotes against source text." </example>

Verify review synthesis accuracy against raw source reviews. <example> Context: QC coordinator dispatches fidelity check after review synthesis assistant: "I'll use the review fidelity checker to verify synthesis claims against raw reviews." </example>

Synthesize book reviews from multiple sources into a meta-review. <example> Context: Raw reviews have been fetched assistant: "I'll use the reviewer agent to synthesize the collected reviews." </example>

Generate multilingual chapter and book summaries. <example> Context: Chunks are ready for a non-fiction book assistant: "I'll use the summarizer agent to create chapter summaries." </example>

Extraction schemas, QC criteria, provider routing rules, and page estimation for the reading-assistant plugin

Obsidian note templates, naming conventions, and wikilink rules for Zettelkasten output