We have published an MCP server for Workerflow. Point any client that speaks the Model Context Protocol at it and your assistant can build and operate automations in your account, using your API key and your permissions.
What it can do
The server exposes the parts of the platform you would otherwise click through:
- Create, update, validate and deploy workflows
- Run a workflow and read back the full execution log
- Create and manage standalone agents, and send them messages
- Create data stores, insert and query records, and export them
- Inspect sessions, executions and usage statistics
In practice this turns "describe the automation you want" into a working draft you can open in the editor and refine.
Getting started
Install the package and add it to your MCP client configuration:
{
"mcpServers": {
"workerflow": {
"command": "npx",
"args": ["-y", "@workerflow/mcp-server"],
"env": {
"WORKERFLOW_API_KEY": "your-api-key"
}
}
}
}
Generate an API key from your account settings, restart your client, and the Workerflow tools appear alongside the rest.
The guardrails still apply
An assistant working through the MCP server is a normal API client. It is scoped to the key you give it, every action it takes lands in the same audit log as everything else, and workflows it creates start as drafts. Nothing runs against production until you deploy it.
The point is not to let a model run your business unattended. It is to remove the mechanical part of building an automation so the review step is where your attention goes.
Full tool reference and setup instructions live in the package README.