home / mcp / backlog mcp server

Backlog MCP Server

Play project for MCP Server Learning

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "danielscholl-backlog-manager-mcp": {
      "url": "http://localhost:8050/sse",
      "headers": {
        "HOST": "0.0.0.0",
        "PORT": "8050",
        "TRANSPORT": "stdio",
        "TASKS_FILE": "tasks.json"
      }
    }
  }
}

Backlog Manager MCP Server offers a lightweight, file-based backlog and issue-tracking system that you can drive with MCP clients. It lets AI agents and other clients create issues, attach tasks, and monitor progress using a simple, portable storage format.

How to use

You connect an MCP client to the server using either SSE (HTTP) or stdio transports. With SSE, you start the server and point your client to the provided HTTP endpoint to send and receive MCP messages. With stdio, you run the server locally and the client communicates directly through standard input/output. Your main workflow is to create issues for features or bugs, add tasks under those issues, and update task statuses as work progresses.

How to install

# Prerequisites
# - Python 3.12 or higher
# - uvx (recommended) or pip
# - Docker (optional for containerized deployment)

# Install with uvx (recommended)
uv pip install -e backlog-manager-mcp

# Verify the installation by starting the server (SSE mode will be shown below)
uv run backlog-manager

# Alternative: run directly with Python (SSE mode)
uv run src/backlog_manager/main.py
````} ,{

Additional sections

Configuration, security considerations, examples, troubleshooting tips, and notes are provided to help you run and integrate the server smoothly.

Configuration

Configure behavior with environment variables. Create a configuration file from the example and tailor the values to your environment.

Security

If you expose the server over a network, ensure you operate behind appropriate access controls and use secure transports where supported. Keep storage files in a location with proper file permissions.

Troubleshooting

Common issues include transport misconfiguration, invalid storage paths, or port binding conflicts. Start by confirming the transport type, storage file path, and port are correctly set, then check startup logs for hints.

Notes

This server is designed to work with MCP clients such as Claude Code or Windsurf. Use the SSE URL for HTTP-based clients or configure stdio for local client integrations.

Available tools

create_issue

Create a new issue with a name and optional description and status.

list_issues

Show all available issues.

select_issue

Set the active issue by name for subsequent task operations.

initialize_issue

Create or reset an issue with an optional description and status.

update_issue_status

Update the status of a named issue.

add_task

Add a task to the active issue with a title and optional description.

list_tasks

List tasks within the active issue, optionally filtered by status.

update_task_status

Update the status of a specific task by its ID.