Side-by-side comparison of two Model Context Protocol servers — install paths, tools, quality signals, and directory engagement so you can pick the right one for Claude, Cursor, and other MCP clients.
Decompose text into classified semantic units with authority, risk, attention scores, and entity extraction. No LLM. Deterministic. Works as MCP server or CLI.
Self-hosted web research for MCP agents: search (SearXNG, with DuckDuckGo fallback), crawl, dense+BM25 rerank, and dedupe into a source-grounded, cited prompt. Local ONNX embeddings by default, or bring an OpenAI-compatible embedding API.
