Index and search desktop files offline with semantic retrieval.
npx skills add https://github.com/kortix-ai/lss --skill kortix-semantic-search
Please help me install this Agent Skill. Skill: kortix-semantic-search Source: https://github.com/kortix-ai/lss/tree/main Command: npx skills add https://github.com/kortix-ai/lss --skill kortix-semantic-search
Locating and recalling relevant documents and contextual knowledge across a desktop workspace can be time-consuming and error-prone when relying on manual file browsing or scattered notes. This Skill provides a local semantic search capability that indexes text across files and agent memory, enabling fast, natural-language retrieval without sending data to the cloud.
Index a directory and start searching with natural language queries.
High-intent search queries and answers about installing and using this skill.
Offline semantic search on local desktop files is done by indexing directories with a local engine that uses BM25 and embeddings, enabling fast natural-language retrieval without internet access.
Yes, you can index and search PDF and DOCX files locally. The semantic search engine supports a wide range of formats including XLSX, PPTX, HTML, JSON, CSV, and code for context-rich retrieval.
Providing agent memory for an AI assistant is best done by indexing contextual knowledge and local desktop files into an offline semantic search engine, allowing the assistant to surface background information quickly.
Yes, local semantic search works entirely offline. It enforces safe, offline operation with configurable embedding providers to ensure your desktop files and agent memory stay on your device.
To keep search results consistent when re-indexing local documents, the semantic search engine uses deterministic chunking and a frontmatter-driven discovery model to maintain index and search alignment.
jiunbaeExtract and convert text, tables, images, and metadata across PDF, DOCX, XLSX, PPTX formats.
ortus-boxlangLoad, chunk, embed, and query documents with vector memory for grounded AI responses.
robwestzSearch an indexed knowledge base with natural language and return semantically ranked results.
InverterNetworkCreate serverless vector indexes and run hybrid search queries.
