FlowCastle ships a built-in MCP server, so AI agents like Claude and Cursor can work on your bot flows directly — reading the current flow, proposing changes, and applying validated edits. You describe what the bot should do; the agent builds it inside your workspace.
MCP (Model Context Protocol) is the open standard that lets AI assistants use external tools. FlowCastle's MCP server exposes your workspace to any MCP-compatible agent: it can inspect your flows and variables, browse the module catalog, look up working flow examples, and apply changes — all through the same validated actions the visual editor uses.
That means you can sit in Claude or your IDE and say "add a lead-qualification branch after the welcome message, and sync the answers to Google Sheets" — and watch the flow update.
Most bot platforms added a chat-style "AI helper" inside their own UI. FlowCastle plugs into the agents you already use. Your prompts, context, and workflow stay where you work — the bot building comes to you.
Connect your MCP client to the FlowCastle MCP server with your workspace credentials, and the tools appear automatically. Works with Claude, Cursor, and any other MCP-compatible client.
A built-in server for the Model Context Protocol — the open standard that lets AI assistants use external tools. It exposes your workspace so an AI agent can inspect flows and variables, browse the module catalog, and apply edits.
Any MCP-compatible agent, including Claude and AI-powered IDEs like Cursor. You describe the change in chat and watch the flow update in your workspace.
Agents work through the same validated actions the visual editor uses, so invalid changes are rejected before they touch your flow.