Glee

Glee is a battery-included MCP toolkit for Claude Code that adds persistent memory, AI code review, session hooks, Git forensics, DB inspection, and background task delegation. Once registered as an MCP server, Claude Code gains access to a suite of tools such as glee.memory.add, glee.memory.search, glee.memory.overview for project memory; glee.code_review for code quality feedback; and glee.config.set/unset for managing configuration. The CLI commands expose memory visibility, code review, and configuration management, while the session hooks ensure relevant project context is available at the start of a Claude session and summarized at the end for memory and traceability. Use glee init claude to register the MCP server and then start Claude Code to begin using the integrated tools.

After setup, you can run memory operations to inspect or augment project memory, invoke code reviews on code or directories, and configure reviewer preferences. The session hooks help Claude Code stay aligned with the project context across sessions, while the Git forensics and DB inspection capabilities provide deeper project insights when debugging or auditing behavior. The toolset is designed to work together, so you can perform end-to-end tasks—from memory recall to code review and session-context management—without leaving your development environment.

Prerequisites:

• Python 3.13 (or a compatible Python runtime)
• Access to the uv tooling ecosystem (uv) for package installation

Install the Glee MCP server:

# Using uv to install Glee (Python runtime)
uv tool install glee --python 3.13

Alternative installation (if you prefer pipx):

pipx install glee-code

Initialize and register the MCP server with Claude Code:

glee init claude

Restart Claude Code to load the new MCP server integration. You should then see glee tools available in your Claude Code session and MCP interface.

Tips and common considerations:

• The MCP server is registered under the name glee; you can customize or alias as needed during setup.
• Tools include memory.add, memory.search, memory.overview, code_review, status, config.set, and config.unset.
• Ensure your environment has access to persistent storage for memory across sessions; the .glee directory and .mcp.json are created during setup.
• If you run into issues starting Claude Code with the MCP server, verify that the uv installation path and the glee package are accessible to the runtime, and confirm that the MCP registration file (.mcp.json) is present in the project root.
• Review the documentation for the Full MCP Service List (GleeMCP/Glee) to understand the roadmap and any breaking changes when updating versions.