Fetch (web page retrieval)

Lets an AI assistant fetch and read the actual content of a URL you give it — useful for having it summarize an article, check current documentation, or reference a specific page instead of relying on its training data, which can be outdated or simply not have seen that page. ```json { "mcpServers": { "fetch": { "command": "uvx", "args": ["mcp-server-fetch"] } } } ``` This only fetches pages you (or the assistant, acting on your request) explicitly point it at — it's not a general web search or autonomous browsing tool. Pair it with a search-focused MCP server if you also want the assistant to find URLs on its own rather than being given one directly. Some sites block automated fetches or require JavaScript rendering the fetcher won't execute — expect it to fail or return incomplete content on those, not every page will work.