Docs.rs

Official Rust SDK for embedding MicroClaw agents, Skills, and Workers

GPUI global-state integration helpers for es-fluent

Polite crawling for pagesift: RFC 9309 robots.txt, per-host rate limits, adaptive backoff and unified error codes.

Pluggable web search backends (SearXNG, Brave, Bing, Google CSE, Tavily) behind one trait.

Process-level tuning for the pagesift binary (glibc allocator thresholds).

PDF to clean Markdown: text extraction, layout reconstruction and scanned-page detection, built for LLM pipelines.

Self-contained MCP server (JSON-RPC 2.0, stdio and Streamable HTTP) for pagesift's scraping tools.

OpenAI-compatible Chat Completions client with structured extraction and summarization.

Block-level Markdown snapshot diff that separates noise changes (price, date, count) from substantive ones.

Schema-based structured extraction (CSS/XPath), regex entity extraction, Markdown chunking and BM25 filtering for pagesift.

HTML to clean Markdown: main-content extraction, Markdown generation and metadata, built for LLM pipelines.

A small, self-contained Chrome DevTools Protocol client and bounded browser pool for rendering pages.

Firecrawl v2 compatible HTTP API for pagesift.

Fast HTML parser for content extraction: zero-copy arena DOM with the HTML5 tree-building rules that matter for text.

Channel-independent Agent Engine and default services for MicroClaw