MCP Toolz

You can run a local MCP server to enable cross-session context saving, todo tracking, and AI-assisted feedback directly from your coding environment. This MCP server integrates with Claude Code to provide persistent storage, session tracking, full-text search, and tools to save, query, and compare contexts and todos across sessions and AI models.

How to use

You connect to the MCP server from your MCP client (such as Claude Code) by configuring an MCP server entry. The server exposes a set of context, todo, and AI feedback tools that let you save items, search across sessions, and get insights from multiple AI models. Use the provided commands to save contexts, run analyses, and restore your todos across sessions. When you save contexts, they are automatically tagged with your current project path, session ID, and session start time, so you can locate and restore work from past sessions.

How to install

Prerequisites: you need Python and a working internet connection. You will also install or clone the MCP Toolz package so you can run the MCP server locally or via a pip installation.

# Install from PyPI (recommended)
pip install mcp-toolz

# Or install from source (development mode)
git clone https://github.com/taylorleese/mcp-toolz.git
cd mcp-toolz
python3 -m venv venv
source venv/bin/activate  # macOS/Linux
# or: venv\Scripts\activate  # Windows

# Install in editable mode with dev dependencies
pip install -e ".[dev]"

MCP Server setup and configuration

You configure the MCP server in Claude Code settings to run the MCP server process. Two explicit runtime configurations are provided depending on how you install the package.

{
  "mcpServers": {
    "mcp-toolz": {
      "command": "python",
      "args": ["-m", "mcp_server"],
      "env": {
        "OPENAI_API_KEY": "sk-...",
        "ANTHROPIC_API_KEY": "sk-ant-...",
        "GOOGLE_API_KEY": "...",
        "DEEPSEEK_API_KEY": "sk-..."
      }
    }
  }
}

MCP Server setup from source

If you installed from source, include the working directory and Python path so the server can locate its modules.

{
  "mcpServers": {
    "mcp-toolz": {
      "command": "python",
      "args": ["-m", "mcp_server"],
      "cwd": "/absolute/path/to/mcp-toolz",
      "env": {
        "PYTHONPATH": "/absolute/path/to/mcp-toolz/src"
      }
    }
  }
}

Available tools

context_save

Save a new context, automatically including session info such as project path and session details.

context_search

Search contexts by query or tags across sessions.

context_get

Retrieve a context by its ID.

context_list

List recent contexts for quick review.

context_delete

Delete a context by its ID.

ask_chatgpt

Request ChatGPT's analysis of a saved context.

ask_claude

Request Claude's analysis of a saved context.

ask_gemini

Request Gemini's analysis of a saved context.

ask_deepseek

Request DeepSeek's analysis of a saved context.

todo_search

Search across saved todos.

todo_get

Get a todo snapshot by its ID.

todo_list

List recent todo snapshots.

todo_save

Save a new todo snapshot.

todo_restore

Restore the active or a specific todo snapshot.

todo_delete

Delete a todo snapshot by its ID.