Convert PDFs, URLs, or text into memory-ready Q&A pairs.
npx skills add https://github.com/grahama1970/fetcher --skill distill-grahama1970
Please help me install this Agent Skill. Skill: distill Source: https://github.com/grahama1970/fetcher/tree/main/.agents/skills/distill Command: npx skills add https://github.com/grahama1970/fetcher --skill distill-grahama1970
Distill content from PDFs, URLs, or plain text into concise, memory-friendly Q&A pairs that can be stored and recalled on demand.
Distill a PDF, URL, or text into memory with domain context to generate Q&A pairs.
High-intent search queries and answers about installing and using this skill.
You can distill web page content into memory-ready Q&A pairs by providing the URL and an optional domain context. The system extracts the text, grounds the answers in the source material, and stores the Q&A pairs in memory.
Yes, you can tailor extracted Q&A pairs by providing a domain context parameter such as "ML researcher." This context shapes the distillation process to generate memory-ready Q&A pairs relevant to your specific field.
The distillation process enforces grounding by deriving answers directly from the provided source content. It uses an LLM as needed to ensure the generated Q&A pairs accurately reflect the original PDF, web page, or text.
The best way to convert plain text into memory-ready Q&A pairs is using a distillation workflow that accepts raw text and domain context. It generates grounded Q&A pairs and persists them via a memory-agent learn workflow for quick recall.
Distilled Q&A pairs are stored in a memory system using the memory-agent learn workflow. This persistence allows the generated Q&A pairs to be quickly recalled on demand for future reference.
decocmsExtract embedded text from PDF files page by page.
doravidanQuery and manage RALPH's persistent memory stored in .ralph/memory/*.json files.
duanbiao2000Build knowledge-grounded LLM applications with retrieval-augmented generation.
unfinishablemapAnalyze site content for logical gaps and unsupported claims.
