v0.4.2

Fixed

• Discover links from rendered HTML when using hosted markdown pages, preserving sidebar/index navigation that .md pages omit.

v0.4.1

Changed

• Update README to emphasize the llms.txt-first workflow and show mirrored output under a references/ directory.

v0.4.0

Added

• Prefer hosted LLM-friendly markdown pages (.md, .markdown, .txt) before falling back to rendered HTML extraction.
• Probe llms.txt indexes and seed crawls from their same-domain markdown links when available.

Fixed

• Report the package version from llm-docs --version and reserve -v for version output.

v0.3.0

Added

• llm-docs cache --site <hostname> to clear cached pages for a specific site.

Changed

• Group cache entries by hostname under ~/.cache/llm-docs.
• Discourage --name for normal website scrapes in CLI help, runtime output, and agent-facing README guidance.

0.2.0

Changed

• Breaking: --scope renamed to --restrict-path-override to discourage overuse. The crawler's prefix-priority system already handles most cases without a hard path restriction.

v0.1.0

Initial release.

Added

• Prefix-priority crawler with depth and URL limits
• Playwright-based JS rendering (headless Chromium)
• Turndown HTML-to-markdown conversion with DOM cleanup
• Site-specific vendor rules (Shopify, Microsoft Learn)
• File-based cache with 7-day TTL
• robots.txt support (on by default)
• links subcommand for discovering unscraped pages
• links --fix to rewrite absolute URLs to relative paths
• --scope flag and auto-scoping site profiles
• --include / --exclude path and regex filters
• --keep-query-strings for versioned doc sites
• --name for custom output directory names

Install:

npm install -g llm-docs-cli
npx playwright install chromium

Full Changelog: https://github.com/mparq/llm-docs-cli/commits/v0.1.0