What problem does it solve? Paid web research tools like Firecrawl stop working without credits, and scattered single-purpose search scripts are hard to maintain. This Skill consolidates web search, JavaScript-rendered page fetching, arXiv search, and Semantic Scholar citation lookups into one local MCP server that requires no API keys, no Docker, and no paid credits. ## Core Features & Use Cases - Keyless Web Search & Fetch: Search via DuckDuckGo (ddgs) and render JavaScript-heavy SPA pages with Playwright where curl or trafilatura return empty shells. - Academic Research: Query the arXiv Atom API for papers with titles, authors, and abstracts, then look up citation statistics through Semantic Scholar. - Use Case: You need to research a topic for a technical blog post. Use web_search to find sources, web_fetch to read JS-rendered documentation pages, and arxiv_search plus scholar_lookup to gather and rank relevant academic papers by citation count. ## Quick Start Register the research MCP server with hermes mcp add using the Hermes venv Python and the research_mcp_server.py path, then restart Hermes and ask the agent to search the web for a topic.
