SV MCP

SV MCP is a FastMCP server from SEO Vendor (SV) that exposes the company's SEO and GEO tools natively to AI agents such as Claude Desktop, Claude Code, Cursor and other MCP hosts. It is built on the sv-cli core library, so resolver, auth and async-task logic match the CLI and REST API. Hosts connect to a single base URL over Streamable HTTP.

What It Is

SV MCP is a Model Context Protocol server that wraps 15 SV API tools as MCP tools, plus get_task_status and get_task_result for async work. SEO Image is available through the API and CLI only. The tools cover short-form and long-form content generation, content transformation, keyword research, entity and topic discovery, topical authority planning, on-page analysis, quick audits, GEO visibility audits, competitor comparison, keyword-to-page mapping, AI-engine rank tracking, content quality scoring, top competitor lookup and a marketplace services search.

How Async Tasks and Calls Work

When an agent calls a tool, the server delegates to the sv-cli core and returns structured JSON. GEO Audit, Prose, SEO Strategist and SEO Mapping are async and return a task ID immediately. The agent can pass wait: true to wait up to 45 seconds, then follow up with the task status and result tools. The SV API allows 1 request per second per API key, and the MCP server retries automatically if that limit is reached.

Connecting and Authentication

For Claude.ai and Claude Desktop, users add a custom connector with the URL https://mcp.seovendor.co and sign in to SV. Claude Code can add it with claude mcp add or a .mcp.json file. Documented configurations also cover Cursor, VS Code, Windsurf, Cline/Roo and Continue, and stdio-only clients can bridge through mcp-remote. Hosted connections use OAuth 2.1; the SV API key is resolved server-side, cached in memory for 5 minutes and never exposed to the host or model.

Running Locally

SV MCP can also be installed locally with pip install sv-mcp and run through uvx with an SV_API_KEY environment variable.