$ loading_
Provide safe self-hosted web grounding for AI agents and crawlers.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "Groundhog" MCP server from askskill: Run: claude mcp add 'io-github-dmytrome-groundhog-mcp' -- uvx groundhog-mcp
Developers or researchers can use it to provide web evidence for an AI assistant when up-to-date online information is needed. This helps ground answers in actually visited page content.
When simple scraping methods cannot reliably fetch content, its browser-based approach can be used for web crawling. It fits automation flows that need to access and extract webpage content.
Teams that need control over data and deployment can use its self-hosted model to build an internal web grounding service. It is suitable for connecting web retrieval to enterprise AI or agent systems.
It is an MCP tool that provides safe, self-hosted web grounding for AI agents and crawlers, using a browser-based approach to access web content.
Based on the description, it emphasizes being self-hosted, which suggests running it in your own environment. See the source repository for exact deployment steps.
Its focus is not a basic search API, but web grounding for AI agents and crawlers through a stealth browser. For more detailed differences, see the source repository.
Enable AI agents to search the web and extract useful page content.
Give AI agents controlled web access, search, and structured data extraction.
Self-hosted web search, extraction, and cited RAG backend for AI agents.
Enable self-hosted deep web research for AI agents without API keys.
Connect AI coding assistants to your trusted local knowledge sources.
Search, crawl, scrape, and extract web data from multiple sources.