ByteChef LogoByteChef

MCP Server

Connect AI assistants to ByteChef using the Model Context Protocol

MCP Server

ByteChef exposes a Model Context Protocol (MCP) server that lets AI assistants — such as Claude, Cursor, and Windsurf — interact directly with your ByteChef instance to manage projects, workflows, and components.

Get Your MCP Server URL

  1. Open ByteChef and go to SettingsMCP Server

  2. Copy the MCP server URL — it has the form:

    https://<your-host>/api/management/<secret-key>/mcp

Keep this URL private. It grants full access to your projects and workflows. Only share it with trusted applications. You can regenerate the secret key at any time from the same settings page.


Connect an AI Assistant

Claude

Claude.ai (web)

Requires a Claude Pro subscription.

  1. Go to claude.ai/settings/integrations
  2. Click Add More
  3. Paste your MCP server URL and save

Claude Desktop

  1. Open Claude Desktop

  2. Go to SettingsDeveloperEdit ConfigOpen claude_desktop_config.json

  3. Add the following, replacing the URL with yours:

    {
      "mcpServers": {
        "ByteChef": {
          "command": "npx",
          "args": [
            "-y",
            "supergateway",
            "--streamableHttp",
            "https://<your-host>/api/management/<secret-key>/mcp"
          ]
        }
      }
    }
  4. Save the file, then quit and restart Claude Desktop


Cursor

  1. Open Cursor → SettingsCursor SettingsMCPAdd new global MCP server

  2. Paste the following, replacing the URL with yours:

    {
      "mcpServers": {
        "ByteChef": {
          "url": "https://<your-host>/api/management/<secret-key>/mcp"
        }
      }
    }
  3. Save


Windsurf

  1. Open Windsurf → SettingsAdvancedCascadeAdd ServerAdd custom server

  2. Paste the following, replacing the URL with yours:

    {
      "mcpServers": {
        "ByteChef": {
          "url": "https://<your-host>/api/management/<secret-key>/mcp"
        }
      }
    }
  3. Save


Other Clients

Any MCP-compatible client that supports Streamable HTTP transport can connect using the raw server URL:

https://<your-host>/api/management/<secret-key>/mcp

Available Tools

The MCP server exposes the following tools to connected AI assistants.

Projects

ToolDescription
createProjectCreate a new project
updateProjectUpdate an existing project
deleteProjectDelete a project
getProjectGet project details
listProjectsList all projects
searchProjectsSearch projects by keyword
publishProjectPublish a project
getProjectStatusGet the current status of a project

Workflows

ToolDescription
createProjectWorkflowCreate a new workflow in a project
updateWorkflowUpdate an existing workflow
deleteWorkflowDelete a workflow
getWorkflowGet workflow details
listWorkflowsList all workflows
searchWorkflowsSearch workflows by keyword
validateWorkflowValidate a workflow definition
getWorkflowBuildInstructionsGet instructions for building a workflow

Components & Actions

ToolDescription
listComponentsList all available components
getComponentGet component details
searchComponentsSearch components by keyword
listActionsList actions for a component
getActionGet action details
searchActionsSearch actions by keyword
getActionDefinitionGet the full definition of an action
getOutputPropertyGet output property details

Triggers

ToolDescription
listTriggersList triggers for a component
getTriggerGet trigger details
searchTriggersSearch triggers by keyword
getTriggerDefinitionGet the full definition of a trigger

Tasks

ToolDescription
listTasksList available task dispatchers
getTaskGet task dispatcher details
searchTasksSearch tasks by keyword
validateTaskValidate a task configuration
getTaskDefinitionGet the full definition of a task
getTaskDispatcherBuildInstructionsGet build instructions for a task dispatcher
getTaskPropertiesGet properties for a task
getTaskOutputPropertyGet the output property of a task
getPropertiesGet general properties

How is this guide?

Last updated on

On this page