wtrl_mcp
Assist LLM writing Waterloo Docstrings, a format with emphasis on machine-verifyable normativity.
README
PRELIMINARY - OFFICIAL START AT 2026-07-01
<p align="center"> <img src="img/wtrl_logo_color.svg" alt="Waterloo Logo" width="240"> </p>
README
Install from source
package_main contains the Waterloo Python package and the Sphinx extension.
The package can be installed from a local checkout of this repository:
cd package_main
pip install .
If you want to build the documentation with the Sphinx extension, install
the optional extra that provides the external sphinx dependency:
cd package_main
pip install ".[sphinx]"
The package can also be installed directly from GitHub:
pip install "git+https://github.com/uwe-at-sdv/sdv_doc_waterloo.git@main#subdirectory=package_main"
SSH works as well for authenticated access:
pip install "git+ssh://git@github.com/uwe-at-sdv/sdv_doc_waterloo.git@main#subdirectory=package_main"
Quick tutorials
waterlint
waterlint is the main command-line tool for Waterloo Docstrings.
It validates docstrings, renders Waterloo JSON, generates helper output, and
supports the MCP-related workflows in this project.
Its JSON diagnostics are machine-readable and therefore well suited for CI
pipelines and GitHub Workflows.
Start with the general help:
waterlint -h
For topic-specific help, use for example:
waterlint help --topic validate
waterlint help --topic render-json
waterlint help --topic render-docker
One important waterlint output path is the MCP server setup: the tool can
validate the server configuration and generate a Docker build script when you
want to package the server as an image.
Running the MCP-server
The MCP server exposes Waterloo documentation data through the Model Context Protocol. It is tightly bound to the Waterloo JSON schema and provides tools that help generate and inspect Waterloo docstrings. It serves the bundled roots, object details, examples, references, and prompt templates so clients can inspect the documentation structure without parsing the JSON documents directly.
After installation, start the server in a terminal with:
wtrl_mcp
This launches the MCP server with the default configuration
src/sdv/doc/waterloo/mcp/wtrl_mcp.toml for access to 127.0.0.1 or
localhost on port 13316.
Consult your operating system documentation if you want to run it as a service.
Registering the MCP-server in Codex or Claude Code
Once the server is running, register it in your CLI client so it can reach the server without manual setup every time. The exact subcommand syntax depends on the CLI version, but the pattern is the same:
codex
codex mcp add myserver --url http://127.0.0.1:13316/mcp
claude
claude mcp add --transport http myserver http://127.0.0.1:13316/mcp
If you prefer localhost, use that host name instead of 127.0.0.1.
If the agent itself runs inside a container or on another host, replace the
address with the one that matches your deployment. The server name is only an
identifier for the client, so you can choose any name that is convenient for
you.
Rendering a Docker setup from the MCP configuration
If you want to ship the MCP server as a container, waterlint can generate a
Docker build script from the MCP configuration. For example:
waterlint render-docker --in src/sdv/doc/waterloo/mcp/wtrl_mcp.toml --out /tmp/myserver.docker
The generated script contains the Docker build steps and the runtime command line. It also prints a few practical hints, such as which port mapping to use and which host names should be allowed for the resulting server. After that, you can build the image with the generated script and run it in Docker.
Projects using Waterloo Docstrings
The following projects use Waterloo Docstrings in practice.
Human-readable HTML
3DE4 Python Scripting Interface
Interactive HTML documentation.
LLM-ready JSON
3DE4 Python Scripting Interface
Waterloo JSON documentation with examples.
Sitemap
Public documentation
- Human-readable documentation
- Example for an interactive presentation
- Example for LLM-ready documentation, best consumed either directly by a coding agent or through an MCP server
Core package
- JSON-Schema
src/sdv/doc/waterloo/schema/wtrl-*-json-*.*.*.schema.json
- Waterloo parser and linter source code
src/sdv/doc/waterloo
- Documentation source (reST)
src/sdv/doc/waterloo/doc
- MCP configuration and server code
src/sdv/doc/waterloo/mcp
- Pytests orchestration and sample code
src/sdv/doc/waterloo/pytestsrc/sdv/doc/waterloo/examples-pythonsrc/sdv/doc/waterloo/examples-diagnostics-python
- Images for public presentation (logos)
src/sdv/doc/waterloo/img
- Tools (not well-documented)
src/sdv/doc/waterloo/tools
IDE extras
- Additional utilities
- Clone directly:
git clone --branch ide-plugins --single-branch https://github.com/uwe-at-sdv/sdv_doc_waterloo.git
- Lexer for
pygmentspygments/python_waterloo_lexer.py
- Extension for
vscode- Waterloo syntax highlighting
- Context menu commands for docstring generation and validation
- Clone directly:
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.