Doc-Scraper Turns Documentation Sites Into Offline Markdown Corpora So Coding Agents Stop Paying for Retrieval
doc-scraper on GitHub, via Hacker News Show HN (single source)·low signal
Doc-scraper hit Show HN on 2026-09-08 (Apache-2.0, 98 stars, last pushed 2026-09-06): a Go crawler that walks a documentation site and emits clean Markdown for LLM ingestion, aimed at building a searchable local corpus for coding agents rather than a RAG service. It is a small tool but it sits on the same line as Archprint and Airuncode from the same 48 hours, all of which replace a hosted context or verification service with something that runs on the developer's machine. Single-source and early-stage, so treat the traction figure rather than the pitch as the signal.