Structure long-form documents into sentence-addressable HTML with semantic search.
npx skills add https://github.com/Talegorithm/Glynk --skill glynk
Please help me install this Agent Skill. Skill: glynk Source: https://github.com/Talegorithm/Glynk/tree/main Command: npx skills add https://github.com/Talegorithm/Glynk --skill glynk
Glynk solves the problem of information overload by transforming messy long-form content like books, papers, and podcasts into structured, sentence-addressable knowledge that both humans and AI agents can effectively read, search, and annotate.
Use the glynk skill to ingest the article at the provided URL and generate a summary of its key takeaways.
High-intent search queries and answers about installing and using this skill.
To structure long-form documents for semantic search, you can ingest PDFs, EPUBs, and web articles into a unified HTML format with sentence-level IDs. This enables precise semantic retrieval and annotation across diverse content.
Yes, you can ingest EPUBs, PDFs, and web articles into a unified HTML format. It provides a token-efficient simplified view for AI agents and a full-featured rendered view for human readers.
pgvector enables semantic search by storing document embeddings in PostgreSQL. This allows high-precision semantic retrieval across all ingested content and user-generated annotations like highlights and summaries.
Yes, PostgreSQL with pgvector is required. It provides the underlying vector storage infrastructure needed to perform high-precision semantic retrieval across your structured documents and annotations.
The best way to structure transcripts is ingesting them into sentence-addressable HTML. This facilitates AI-assisted reading workflows by providing unified API access to the ingested content and user annotations.
Ingesting diverse formats into sentence-addressable HTML requires resolving structural complexities across PDFs, EPUBs, and web articles. Complex layouts may need preprocessing to ensure accurate sentence-level ID assignment and semantic retrieval.
markus41Manage vector embeddings for semantic search across Pinecone, Weaviate, Chroma, and pgvector.
oaustegardGenerate interactive HTML demo pages with bookmarklet functionality.
bbartlingExpose open-fdd site, rule, data, and visualization functions via FastAPI REST endpoints.
ancolemanImplement vector databases for AI/ML semantic search and RAG systems.
