Browse
→ Web & Browser
→ WebVector MCP
WebVector MCP
MCP server, CLI and TypeScript library for web research in one tool call: search, fetch full pages, chunk, rank and return cited passages. It works with zero API keys using lightweight local retrieval, with optional embeddings from local Transformers or external providers when configured. Public GitHub repository created August 18, 2026.
MCP unverified
Integration
| Transport | stdio |
| Auth | none |
| Endpoint | https://github.com/rthomas24/web-vector |
| Install | npx -y webvector-mcp |
Use Cases
| 01 | Let an assistant perform cited web research by searching, reading and ranking pages in a single MCP call |
| 02 | Fetch HTML, PDFs or Markdown and return evidence-bearing passages without requiring an API key |
| 03 | Upgrade ranking to hybrid semantic retrieval when local or hosted embedding providers are available |
Tags
web-research search rag citations vector-search duckduckgo embeddings typescript
Machine-readable: /api/servers.json
· JSON-LD schema embedded in <head>