toro4n
使用 AI 助手在 toro4n 中绘制、塑造和编辑 Neo4j 图数据模型。
托管 MCP 服务器
npx add-mcp 'https://toro4n.com/mcp'可安装到 Claude Code、Codex、Cursor 等客户端
文档
Connect the AI assistant you already use, and it can turn a description, a document, a spreadsheet, or a database schema into a graph data model. The model lands in your toro4n account, laid out and ready to edit. Ask for changes while you have it open, and you'll see them as they happen. toro4n works with any assistant that can connect to a remote Model Context Protocol (MCP) server.
Server address
https://toro4n.com/mcp
Assistants work only with models you've saved online, so you'll need to sign in to toro4n, with Google or with a code we email you.
Connect your assistant
- Claude (claude.ai and the desktop and mobile apps): go to Customize › Connectors › Add custom connector. Paste the address, add the connector, and choose Connect. Then, in any chat, use the + button to turn it on.
- ChatGPT: in Settings, under Apps, turn on developer mode. Then create an app, paste the address, and choose OAuth for authentication. Developer mode isn't available on every ChatGPT plan.
- Claude Code:
claude mcp add --transport http toro4n https://toro4n.com/mcp - Antigravity CLI, which replaced Gemini CLI: add toro4n to
~/.gemini/config/mcp_config.json:{"mcpServers": {"toro4n": {"serverUrl": "https://toro4n.com/mcp"}}} - VS Code with Copilot, Cursor, and other MCP clients: add the address as a remote HTTP MCP server.
The first time you connect, your assistant sends you to toro4n to sign in and click Allow. It can't reach your models until you do.
What to ask
“Sketch a graph model of our order system from this document.” “Draw the schema in this spreadsheet as a Neo4j model.” “Add invoices to my bookstore model.” The assistant replies with a link that opens the model in toro4n.
What your assistant can do
get_modeling_guide– toro4n's guide to Neo4j graph data modeling: naming conventions, and when to use a node, a property, a label, a relationship, or an intermediate node. toro4n won't let your assistant create or edit a model until it has read the guide.list_models– the models you've saved online and the ones shared with you, with your role on each and any folder you've filed it in.get_model– a model's nodes, relationships, notes, and layers.create_model– a new model from nodes, relationships, and notes, laid out automatically.edit_model– add, change, or remove nodes, relationships, properties, and notes, or rename and remove layers. All the changes in a call apply together, or not at all.list_versionsandrestore_version– list a model's earlier versions and restore one.
The assistant plays by the same rules you do: each node needs a unique label or combination of labels, names follow Neo4j's naming conventions, and property types come from the Type list. If a change breaks a rule, toro4n tells the assistant what to fix. And it can only do what your role on a model allows.
Staying in control
Before an assistant changes a model, we save a version, so you can go back if you don't like the result. The assistant's changes appear as “you@example.com via Claude” in Version history. To see which assistants are connected, or to disconnect one, go to Account › AI assistants.
Anything your assistant reads from your models goes to the company that runs it, under that company's terms. toro4n never sends your models to an AI provider on its own.
Questions
Which AI assistants work with toro4n?
Any assistant that can connect to a remote MCP server: Claude (claude.ai and its apps), ChatGPT in developer mode, Claude Code, Antigravity, VS Code with Copilot, Cursor, and others. Connect it to https://toro4n.com/mcp.
Can the assistant delete my models?
No. It can list, read, create, and edit models, and restore earlier versions, but only where your role allows. It can't delete or share a model. If you don't like a change it made, open Version history and restore the version from just before it.