Blowsh
Enables AI agents and clients to browse and extract content from modern, JavaScript-rendered web pages as plain text, HTML, or Markdown.
About
Blowsh is a Model Context Protocol (MCP) server that empowers AI agents, IDE agents, and MCP clients with JavaScript-capable web browsing directly from the terminal. By integrating with Browsh, it allows AI to fetch and render any modern web page, including those with complex JavaScript, and receive the processed content in easily-consumable formats like plain text, structured HTML, or clean Markdown. This facilitates advanced web interaction for AI workflows, enabling summarization, data extraction, and content ingestion from dynamic websites.
Key Features
- fetch_plain Tool: Extracts readable plain text after full JavaScript rendering, ideal for summarization.
- AI-optimized tool documentation: Includes inputs, outputs, and use-cases designed for seamless agent automation.
- 0 GitHub stars
- Robust Browsh management: Manages Browsh lifecycle, ensuring it launches once, stays running, and shuts down gracefully.
- fetch_html Tool: Provides post-JavaScript, real DOM/HTML output for detailed parsing and scraping.
- fetch_markdown Tool: Converts JS-rendered HTML into clean Markdown for optimal LLM context and semantic search.
Use Cases
- Extracting readable plain text from dynamic websites for summarization or natural language processing tasks.
- Parsing after-JavaScript rendered HTML for web scraping, data table extraction, or link graph construction.
- Generating Markdown-formatted content for large language model context ingestion, semantic search, or retrieval-augmented generation.