Rework

You can run a Model Context Protocol (MCP) server that lets AI agents interact with Rework projects, tasks, workflows, and jobs. This server provides a standardized way to manage your work items through natural language commands and scripted operations, making it simpler to orchestrate complex AI-assisted workflows.

How to use

Start the server with your Rework credentials and run it in your preferred mode. You can operate locally via a standard MCP client to create tasks, fetch tasks, manage workflows, and track jobs within Rework. You can also enable Server-Sent Events (SSE) for real-time updates and status streaming. When you issue commands to the MCP client, you will be able to manage projects, assign users, and monitor progress using natural language prompts.

Key capabilities you can leverage include creating tasks, listing tasks in boards, viewing task details, adding comments, managing projects, creating and tracking jobs, listing jobs by workflow, and linking tasks to users. You can also search for users by name or email, assign users to tasks, and adjust user permissions as needed.

How to install

version: '3.8'

services:
  rework-mcp-server:
    build:
      context: .
      dockerfile: Dockerfile
    ports:
      - '3000:8000'
    environment:
      - REWORK_PROJECT_ACCESS_TOKEN=${REWORK_PROJECT_ACCESS_TOKEN}
      - REWORK_PROJECT_PASSWORD=${REWORK_PROJECT_PASSWORD}
      - REWORK_WORKFLOW_ACCESS_TOKEN=${REWORK_WORKFLOW_ACCESS_TOKEN}
      - REWORK_WORKFLOW_PASSWORD=${REWORK_WORKFLOW_PASSWORD}
      - REWORK_ACCOUNT_ACCESS_TOKEN=${REWORK_ACCOUNT_ACCESS_TOKEN}
      - REWORK_ACCOUNT_PASSWORD=${REWORK_ACCOUNT_PASSWORD}
      - ENABLE_SSE=true
      - LOG_LEVEL=info
    volumes:
      - ./src:/app/src
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 5s

Additional configuration and notes

Prerequisites: you need Node.js installed to run the MCP server directly or you can run it via a package runner. You will also configure environment variables with your Rework access tokens and passwords for projects, workflows, and accounts.

You can enable Server-Sent Events (SSE) to receive live updates by setting ENABLE_SSE to true and choosing a port (default is 3000 if not specified). SSE is useful for real-time progress and status streams during long-running workflows.

If you want to disable certain tools to simplify the available actions, set the DISABLED_TOOLS environment variable to a comma-separated list of tool names such as create_task,get_tasks. This helps manage context length and reduces complexity.

Custom fields are supported for tasks and jobs through the custom_fields parameter, which lets you specify field identifiers and values to carry extra metadata like priority or custom attributes.

Troubleshooting and tips

Ensure all required environment variables are provided and correctly named. If you encounter authentication or permission issues, double-check the access tokens and passwords you supplied for each Rework area (projects, workflows, accounts).

If you see rate limiting or API errors, review your token scopes and make sure you are not hitting API usage quotas. Review LOG_LEVEL to debug issues in the server logs.

Available tools

create_task

Create a new task in a board with a name, associated project, and creator information.

get_tasks

Retrieve tasks from a board with optional filters to narrow results.

get_detail_task

Fetch detailed information for a specific task by its ID.

get_projects

List all projects and boards with an optional search query.

create_job

Create a new job within a workflow, specifying its name and creator.

get_jobs

Retrieve jobs from a workflow with optional filters.

get_detail_job

Fetch detailed information for a specific job by its ID.

get_workflows

List all available workflows with an optional search query.

find_users

Find users by name or email with optional properties filtering.

list_all_users

List all workspace members.