How to switch AI coding assistants without starting over

If you're moving to a different AI coding assistant, this Graphify guide shows how to keep your code knowledge graph instead of rebuilding it. The graph lives in your Graphify account, not inside any one tool, so a new client signs in to the same account, picks the workspace that holds your repositories, and reads the same graph.

Every client connects to the same hosted MCP server, https://api.graphify.com/mcp, which uses Streamable HTTP with OAuth. You sign in once in the browser and there's no API key to paste. The guide lists the connection step for each assistant:

• Claude Code: claude mcp add --transport http --scope user graphify https://api.graphify.com/mcp, then authenticate from /mcp
• Cursor: a one-click install link, or add the URL to ~/.cursor/mcp.json
• Codex: the Graphify plugin, or codex mcp add graphify --url https://api.graphify.com/mcp and then codex mcp login graphify
• VS Code: the Graphify MCP extension (VS Code 1.102 or later), or an mcp.json entry
• GitHub Copilot CLI: copilot mcp add --transport http graphify https://api.graphify.com/mcp

To check the connection, ask the new assistant to run Graphify's list_repositories tool. The list should match your workspace. Then ask a structural question you already know the answer to, like what calls a function you wrote. The Free plan covers up to 3 repositories with no card.

Graphify doesn't move chat history, settings, or subscriptions, so export those from the old tool yourself. It also only serves what you've pushed, and an assistant may still reach for grep unless your instructions file tells it to use the graph. Copilot on github.com can't connect yet, because its coding agent and code review only support static header credentials. If you're leaving Conductor Cloud, which starts deleting user data on Nov 20, the guide covers Conductor's export routes, but Graphify has no Conductor integration.