Pylpex MCP
# Pylpex MCP
This is a MCP server for Pylpex interpreter.
## Installation
Assuming you are using `Claude Desktop` and have `uv` installed, follow these instructions to set up the Pylpex MCP server:
1. Clone the repository:
```bash
git clone https://github.com/Golto/pylpex-mcp.git
cd pylpex-mcp
```
2. Install dependencies:
```bash
uv sync
```
This command reads the `pyproject.toml` file and installs all the required dependencies.
3. Configure your Claude Desktop settings to use this MCP server:
Go to your Claude Desktop settings, then to the Developer Settings tab, click on "Edit config" and add the following configuration to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"Pylpex": {
"command": "uv",
"args": [
"--directory",
"$ABSOLUTE/PATH/TO/THE/PROJECT/pylpex-mcp",
"run",
"mcp",
"run",
"main.py"
]
}
}
}
```
You may want to replace `uv` with the absolute path to the `uv` executable. To find `uv` location, use `which uv` on Linux/MacOS or `where uv` on Windows. Also, replace `$ABSOLUTE/PATH/TO/THE/PROJECT/pylpex-mcp` with the absolute path to the project directory. `--directory` argument is required for `uv` to find the project with dependencies.
> **Note**: Running `uv run mcp install main.py` will install the MCP server in your Claude Desktop settings directly. But it may not work because of the `--directory` argument missing.
4. Close and reopen your Claude Desktop application to apply the changes.
5. Verify that the Pylpex server is running by checking the status in the Developer Settings tab.
You should see the Pylpex tools available in your Claude Desktop interface.
## Tools
At this point, you should have access to the following tools:
- **run**: Execute Pylpex code and return the result.
- **tokenize**: Convert Pylpex code into tokens for analysis. Use the Lexer component of the Pylpex interpreter.
- **get_variables**: Get all variables in the current interpreter state.
- **reset**: Reset the interpreter state.
## Resources
- **get_documentation**: Get the documentation of the Pylpex language. It gives you the syntax and the available functions.
- **get_examples**: Get examples of Pylpex code as code snippets.
## More details
For more details about the Pylpex language, you can check the [Pylpex github repository](https://github.com/Golto/pylpex).
> **Note**: The Pylpex language is still in development and may change in the future. And it's in french...TDQS
Scored across 4 tools
Each tool has a clearly distinct purpose with no overlap: get_variables retrieves interpreter state, reset clears it, run executes code, and tokenize analyzes code structure. The descriptions make these roles unambiguous, preventing agent misselection.
All tool names follow a consistent verb-based pattern (get_variables, reset, run, tokenize) without mixing conventions. The naming is straightforward and predictable, enhancing readability and agent usability.
With 4 tools, the count is reasonable for a language interpreter server, covering core operations like execution, state management, and analysis. It could benefit from additional tools (e.g., for debugging or syntax validation) but is well-scoped for basic functionality.
The toolset covers essential interpreter workflows: running code, inspecting state, resetting state, and tokenizing code. Minor gaps exist, such as lack of tools for step-through debugging or syntax checking, but agents can work around these with the provided tools.