What it does
The Oxylabs MCP server connects AI models to web content through the Model Context Protocol. Its Web Scraper API tools include universal_scraper for any URL with optional JavaScript rendering and geo-targeting, google_search_scraper, amazon_search_scraper and amazon_product_scraper. Its AI Studio tools include ai_scraper, ai_crawler, ai_browser_agent, ai_search, ai_map and generate_schema. The source states that it can render JavaScript-heavy pages, manage CAPTCHAs and access geo-restricted data from 195+ countries.
An assistant would reach for it when it needs the content of a page, structured search results, Amazon search or product data, data collected across several pages from a starting URL, or a browser that navigates, clicks and fills forms based on a prompt. The two tool sets can be used together or on their own.
There are two ways to run it. The hosted server needs no installation: credentials are sent as request headers, and the server is also listed on Smithery. Locally, it runs through uvx or a clone of the repository, with credentials set as environment variables such as OXYLABS_API_TOKEN or OXYLABS_AI_STUDIO_API_KEY. It can also be self-hosted over the streamable-HTTP transport, where credentials are passed per request. All tools are always listed, and calling one without its credentials returns an error explaining what to configure. Tool-call information is sent as notification/message events.
Limitations and requirements
Requires at least one of: an Oxylabs Web Scraper API account (API token or username and password) or an Oxylabs AI Studio API key. Running locally requires the uv package manager. Only environment variables with real credentials should be set, because placeholder values make tool calls fail. Clients that support only OAuth for remote servers cannot pass the required headers yet. The publisher states that the repository is still supported, but users of the Oxylabs Web API are directed to the separate Web API MCP repository.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
ai.smithery/oxylabs-oxylabs-mcp
Fetch and process content from specified URLs using the Oxylabs Web Scraper API.
v1.13.1As shown on registry.modelcontextprotocol.io when we last read it