Semantic Search & Document Parsing
Parses complex documents and performs high-speed semantic keyword searches across large collections using Rust-powered CLI tools.
The source repository doesn't declare a license. Check its terms before reusing the code.
Key features
- Local semantic keyword search using multilingual embeddings and cosine similarity
- Rust-based CLI utilities designed for high performance and reliability
- High-speed document conversion to markdown using the LlamaParse API
- Persistent workspace management for caching embeddings across large file sets
- Configurable search parameters including top-k results and distance thresholds
Use cases
- Managing and querying large-scale document collections with persistent embedding caches
- Searching through hundreds of research papers or technical manuals for specific concepts
- Converting large batches of proprietary document formats into markdown for LLM processing
FAQ
What does the Semantic Search & Document Parsing skill do?
This skill provides Claude with high-performance Rust-based CLI tools to convert complex documents (PDFs, DOCX, PPTX) into markdown and perform context-aware semantic searches across large file collections using vector embeddings.
Why is semantic search better than standard grep for documentation?
Unlike grep, which only finds exact text matches, semantic search uses multilingual embeddings to find content related to your query's intent. This allows you to find relevant information even if the document uses different terminology than your search terms.
When should I use this skill in my workflow?
Use this skill when you need to reference external documentation, analyze large sets of research papers, or search through project requirements that aren't in plain text. It is ideal for finding information based on meaning rather than just exact keyword matches.
Does this skill require an API key to function?
The parsing component uses the LlamaParse API, which requires a free LLAMA_CLOUD_API_KEY. However, the semantic search and workspace management features run locally on your machine using Rust utilities.
How does workspace management improve performance?
Workspaces allow Claude to cache embeddings for your document collections. By storing these vectors locally, subsequent searches over the same files become nearly instantaneous, as the skill doesn't need to re-process the data every time.
Related skills
Brave Search
Performs privacy-focused web and local business searches to retrieve real-time information, documentation, and news.
Learning Documentation612 ptsAboutMe Index
Accelerates file discovery by mapping project components to their specific purposes using a centralized JSON index.
Learning Documentation112 ptsSmalltalk Usage Finder
Analyzes Pharo Smalltalk codebases to find usage patterns, explain class responsibilities, and discover implementation examples.
Learning Documentation112 pts