web-shooter: a local MCP web reader returning answers at ~1k tokens vs ~25k
Dazzling_Dirt_7729 · reddit · 2026-11-01
A developer released web-shooter, a local MCP server with no built-in LLM and zero storage that feeds only relevant passages to your model instead of full pages. Tools: fetch (pages/PDFs/JSON APIs → ranked passages), links, best-first crawl, and headless-Chrome render. In a 32-question single-run benchmark it hit 29/32 answers at 793ms median latency and 1,081 avg tokens, vs Firecrawl cached (28/32, 25,409 tokens) and Jina Reader (24/32, 25,345 tokens). Caveats: tuned on the test set; no JS-heavy or bot-protected pages. MIT/Apache-2.0 on GitHub.
More from coding & agent
- Reddit debate: is query expansion still needed in RAG when semantic search already matches meaning? — DueBid3298 · 2026-11-01
- Founder builds AI charter-airline agent on Claude handling quotes to invoices — weskambale · 2026-11-01
- Should MCP servers start with zero network? A sandboxed Claude Code setup argues yes — quiet_flare · 2026-11-01
- AI models are getting more agentic — can human oversight actually keep up? — Mountain_Finger4856 · 2026-11-01
- Dev shows off AI-generated game engine material graphs built from text descriptions — D3VAUX · 2026-11-01
- UIUC, Meta, Amazon & DeepMind release 135-page survey charting agentic reasoning for LLMs — mdancho84 · 2026-11-01