backlog -rust

This MCP server exposes the Backlog API to MCP-compatible clients. It allows AI assistants and other tools to read and write data to Backlog, including documents, Git repositories, issues, projects, shared files, users, and wikis. With the default configuration you gain access to a suite of 36 tools organized by category (Documents, Git/Pull Requests, Issues, Projects, Shared Files, Users, and Wikis). Each tool follows a category_resource_action naming pattern (for example, issue_details_get or wiki_update) to enable precise filtering with commands like --allowedTools. You can use this server to fetch information (read operations) and perform actions (write operations) against your Backlog space by authenticating with the provided BACKLOG_API_KEY and pointing the server at your Backlog base URL. The server is designed to be embedded in clients that need Backlog automation within MCP workflows and supports various MCP client configurations (Claude Desktop, Cline, Gemini CLI).

Prerequisites:

• macOS or Linux
• Homebrew installed (recommended)
• Access to a Backlog space with API key permissions

Install using Homebrew (recommended):

# Add the tap (only needed once)
brew tap safx/tap

# Install the Backlog MCP server and optional backlog CLI
brew install mcp-backlog-server  # MCP server for AI assistants
brew install blg                 # Backlog CLI tool (Optional)

Update to the latest version:

brew update
brew upgrade mcp-backlog-server
brew upgrade blg

Uninstall:

brew uninstall mcp-backlog-server
brew uninstall blg
brew untap safx/tap  # Optional: remove the tap

Alternative installation methods:

• Pre-built binaries: Download from the releases page: https://github.com/safx/backlog-mcp-server-rust/releases
• Build from source (Rust):

git clone https://github.com/safx/backlog-mcp-server-rust.git
cd backlog-mcp-server-rust
cargo build --release

If you build from source, the binary will typically be located at target/release/mcp-backlog-server

Notes and tips:

• The server requires these environment variables: BACKLOG_BASE_URL, BACKLOG_API_KEY, BACKLOG_PROJECTS, BACKLOG_PREFIX. Populate them with your space URL, API key, comma-separated project keys, and an optional prefix for virtual Backlog entities.
• The domain constraints are documented in the README: backlog.com, backlog.jp, or backlogtool.com domains are supported.
• Tool names use the category_resource_action pattern to support filtering with --allowedTools in clients such as Claude, ensuring you can narrow down capabilities.
• When downloading attachments or shared files, the MCP server can return content as base64 images, text, or raw bytes based on content type and encoding rules.
• If you encounter authentication or rate-limit issues, verify that BACKLOG_API_KEY has the necessary scopes for the requested operations and that BACKLOG_BASE_URL points to the correct Backlog instance.
• For production deployments, consider running behind a process manager and exposing logs for observability (stdout/stderr).