<div align="center"> <img src="https://raw.githubusercontent.com/liustack/modsearch/main/assets/banner.jpg" width="100%" alt="ModSearch, plug-in web search and page fetch for text-only LLMs" /> <h1>ModSearch</h1> <p><b>Free plug-in web search for your text-only LLM.</b></p> <p> <a href="https://www.npmjs.com/package/@liustack/modsearch"><img src="https://img.shields.io/npm/v/@liustack/modsearch" alt="npm"></a> <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue.svg" alt="License"></a> </p> <p><a href="./README.zh-CN.md">简体中文</a></p> </div>

DeepSeek-V4-Flash gives you a lot of model for very little money: fast and strong, but its built-in web search is weak, and most third-party gateways ship no search at all. A model that cannot look things up or read a web page is a real handicap these days.

ModSearch fixes this with a completely free setup. It never touches your config and never adds a local proxy. It's just a search-engine plug-in, usable as a CLI or as an Agent Skill, that returns structured web search results and can parse pages too. Under the hood it runs on [Antigravity CLI](https://antigravity.google) (`agy`), whose search comes from free-quota Google Search. And Google Search is, without much argument, the best search engine on earth. How it works:

```text your text-only model ──▶ modsearch skill (auto-triggers on fresh-info needs) │ ┌─────────┴─────────┐ ▼ ▼ -q "query" -u <url> web search page fetch └─────────┬─────────┘ ▼ agy · Gemini 3.6 Flash (free quota) │ ▼ structured JSON evidence ──▶ model answers with sources ```

Install the skill once and your agent starts searching and reading the web on its own. No model swap, no API key, no