Analyzes long documents and answers questions via hierarchical index.
npx skills add https://github.com/malue-ai/dazee-small --skill deep-doc-reader
Please help me install this Agent Skill. Skill: deep-doc-reader Source: https://github.com/malue-ai/dazee-small/tree/main/instances/xiaodazi/skills/deep-doc-reader Command: npx skills add https://github.com/malue-ai/dazee-small --skill deep-doc-reader
This Skill enables deep analysis and question-answering for lengthy documents like PDF reports, contracts, and manuals, overcoming the limitations of simple text summarization.
Use the deep-doc-reader skill to analyze the attached document 'annual_report_2023.pdf' and answer the question 'What were the main drivers of revenue growth in Q3?'.
High-intent search queries and answers about installing and using this skill.
To analyze a long PDF report, this Skill builds a hierarchical tree index of sections and subsections, enabling precise section-level retrieval and multi-turn question-answering on specific document content without needing to re-index for follow-up queries.
Yes, deep-doc-reader supports multi-turn Q&A on a single document by constructing a hierarchical tree index once, allowing you to ask sequential follow-up questions about the indexed PDF or manual without reprocessing the entire file.
You can extract specific financial figures from a 100-page PDF contract using precise page-level and section-level retrieval, which locates exact data points within the hierarchical document structure for accurate question answering.
The best way to find information in lengthy PDF manuals is hierarchical indexing, which constructs a tree-like index of sections to enable precise section-level retrieval, delivering accurate answers based on document structure instead of simple text summarization.
Simple text summarization limits long document processing by missing specific details, whereas building a structured tree index overcomes this by enabling precise section-level retrieval and deep reasoning over the original document structure.
Section-level analysis works by utilizing PageIndex MCP to construct a hierarchical tree index of document sections and subsections, allowing the system to locate and reason over precise structural areas for accurate information retrieval.
qte77Index documents into heading-aware chunks with sentence-transformers and FAISS.
jiunbaeExtract and convert text, tables, images, and metadata across PDF, DOCX, XLSX, PPTX formats.
thewoolleymanPerform semantic vector search over documents using AgentDB with HNSW indexing.
takemi-ohamaExtract text, tables, sections, and metadata from PDF documents.
