Linear MCP MCP server

Linear's MCP server gives any compatible AI model or agent access to your Linear workspace. The docs say it has tools for finding, creating and updating objects such as issues, projects and comments, with more functionality on the way. It is hosted and managed by Linear and follows the authenticated remote MCP specification. The server address is https://mcp.linear.app/mcp over Streamable HTTP. Setup uses OAuth 2.1 with dynamic client registration, and you can also pass a bearer token or a Linear API key. For Claude Code the docs give: claude mcp add --transport http linear-server https://mcp.linear.app/mcp, then run /mcp in a session to finish signing in. They also cover the Claude.ai and Claude desktop connectors, Codex, Cursor, Jules, VS Code, v0 by Vercel, Windsurf and Zed, and the mcp-remote module for clients without remote MCP support. We would add it when your work is planned in Linear: the agent can read the issue it has been asked to build, then update it or leave a comment when it is done. The default endpoint is read-write. If you only want the agent to look, connect to https://mcp.linear.app/mcp/readonly, which only exposes read tools, or request just the read OAuth scope. The FAQ also notes that remote MCP connections can occasionally need a reconnect.