Open-source MCP server for AI agents: web search, content extraction, and library docs -- 5-strategy scraping, runs without API keys.
README
mcp-name: io.github.n24q02m/wet-mcp
Web search, content extraction, and library docs for AI agents -- 5-strategy scraping, runs without API keys.
| Phase | Status | Scope |
|---|---|---|
| Phase 1 | Shipped | web-core ScrapingAgent migration, smart chunks output, search polish, media slim |
| Phase 2 | Shipped | Context7-level docs search: library index (Tier 1 + Tier 2), version-aware queries with token cap, project lock (Cabinets) |
| Phase 3 | Shipped | extract.agent multi-step research with cited synthesis, extract.interact click/fill/submit via patchright (optional session persistence), docs_004_chunk_summaries migration, media.analyze removed (v2.0.0) |
<!-- Badge Row 1: Status --> <!-- Badge Row 2: Tech -->Current release: v3.x.
media(action="analyze")was removed in the v2.0.0 BREAKING release. Useimagine-mcp'sunderstandaction for vision/audio/video analysis. Seedocs/migration.mdfor the upgrade recipe.