Fast, local semantic search over web content for AI agents. Hybrid BM25 + potion-retrieval-32M embeddings, cross-page dedup, token-budget mode, MCP server, SearXNG bridge. ~90% fewer tokens than raw web_fetch.
Skill Claude CodeCodex
Token-efficient web research for AI agents. Fetches, indexes, and queries web content locally via hybrid BM25 + embeddings, returning only the chunks that fit a token budget. Replaces raw webfetch for any task involving more than one source. Use when the user asks you to research, compare, summarize, or fact-check…
Skill Claude CodeCodex
Token-efficient web research for AI agents. Fetches, indexes, and queries web content locally via hybrid BM25 + embeddings, returning only the chunks that fit a token budget. Replaces raw webfetch for any task involving more than one source. Use when the user asks you to research, compare, summarize, or fact-check…