1broseidon/ketch

Fast, stateless CLI for web search and scrape. Built for AI agents.

What it solves

Ketch is a stateless CLI tool designed to simplify the process of gathering research data for both humans and AI agents. It eliminates the need to manage multiple provider SDKs, authentication keys, and varying response formats by consolidating web search, code search, library documentation, and web scraping into a single binary.

How it works

Ketch acts as a unified interface for various research surfaces:

  • Web Search: Integrates with providers like Brave, DuckDuckGo, SearXNG, Exa, Firecrawl, Keenable, Tavily, Parallel, and SerpBase.
  • Code Search: Greps open-source software (OSS) source code via Grep, Sourcegraph, or GitHub Code Search.
  • Library Docs: Accesses curated, version-aware documentation via Context7.
  • Scraping/Crawling: Converts HTML pages and text-based PDFs into clean Markdown. It includes a built-in pure-Go PDF parser and supports headless Chrome for JS-rendered pages.

It provides structured JSON output (--json) and documented exit codes, making it highly predictable for scripted control flow and AI agent integration. It also features a local page cache to speed up repeat requests.

Who it’s for

  • AI Agent Developers: Those who want a single tool to provide their agents with research capabilities without writing per-provider glue code.
  • Researchers and Developers: Humans who prefer a fast terminal-based tool for extracting clean content from the web or searching code.

Highlights

  • Unified Interface: One binary for search, code, and docs.
  • Agent-Ready: Structured JSON output, stable exit codes, and an MCP server implementation for direct integration with agents like Claude Code.
  • JS-Rendering Fallback: Automatically detects JS-shell pages and re-fetches using headless Chrome.
  • Rank-Fused Search: The --multi flag allows federated search across multiple backends with Reciprocal Rank Fusion to deduplicate and rank results.
  • PDF Extraction: Built-in text extraction for PDFs with support for external OCR converters.
  • Cookie Support: Ability to use Netscape cookies.txt files to bypass consent walls or access authenticated content.

Related

  • Project
    dondai44423/donsetch

    A high-performance web research tool for AI agents that bypasses bot detection using native Chrome TLS and provides keyless search and token-optimized content extraction.

  • Project
    iOfficeAI/OfficeCLI

    A self-contained CLI tool that allows AI agents and developers to create and edit Word, Excel, and PowerPoint files without needing Microsoft Office installed.

  • Project
    firecrawl/cli

    A command-line interface for Firecrawl that enables searching, scraping, and crawling the web to provide AI-ready data, with specialized indices for developer and research papers.

  • Project
    Aas-ee/open-webSearch

    Open‑WebSearch is a Node‑JS service that lets AI agents perform live web searches and fetch article content without any API keys. It offers an MCP‑compatible server, a CLI, and a long‑running HTTP daemon, aggregates results from many public search engines, supports proxy use, and can optionally use Playwright‑driven headless browsers for sites that need JavaScript rendering.

  • Project
    better-world-ai/x-cli

    A collection of AI-generated command-line interfaces for websites without public APIs, enabling terminal-based interaction with platforms like rental sites and travel booking.