MCP LaTeX Server
Create, edit, and manage LaTeX files. Requires an external LaTeX distribution like MiKTeX, TeX Live, or MacTeX.
MCP LaTeX Server
A Model Context Protocol (MCP) server for LaTeX file creation, editing, validation, and compilation. Built with FastMCP and Pydantic for type-safe, structured output.
Features
- Create LaTeX documents from parameters or bundled templates (article, beamer, report)
- Edit files with replace, insert, append, and prepend operations
- Read and list
.texfiles within a secure base directory - Validate syntax — braces, environments, references, required declarations
- Compile to PDF with
pdflatex,xelatex, orlualatex - Resources — browse and retrieve bundled templates via
latex://URIs
Prerequisites
- Python 3.11.9+
- LaTeX distribution (for compilation):
Installation
Using uv (recommended)
git clone https://github.com/RobertoDure/mcp-latex-server
cd mcp-latex-server
uv pip install -e .
Using pip
git clone <repository-url>
cd mcp-latex-server
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
source .venv/bin/activate
pip install -e .
Quick setup (Windows)
python quick_setup.py
This checks Python version, installs dependencies, verifies the server imports correctly, and optionally configures Claude Desktop.
Configuration
The server uses a single environment variable:
| Variable | Description | Default |
|---|---|---|
LATEX_SERVER_BASE_PATH | Root directory for all file operations | . (current directory) |
All file paths passed to tools are resolved relative to this base directory. Access outside it is denied.
MCP client configuration
Claude Desktop
Add to your claude_desktop_config.json:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
{
"mcpServers": {
"latex-server": {
"command": "uv",
"args": [
"--directory",
"/path/to/mcp-latex-server",
"run",
"latex_server.py"
],
"env": {
"LATEX_SERVER_BASE_PATH": "/path/to/your/latex/files"
}
}
}
}
VS Code (GitHub Copilot)
Add to .vscode/mcp.json in your workspace:
{
"servers": {
"latex-server": {
"command": "uv",
"args": [
"--directory",
"/path/to/mcp-latex-server",
"run",
"latex_server.py"
],
"env": {
"LATEX_SERVER_BASE_PATH": "${workspaceFolder}"
}
}
}
}
Tools
create_latex_file
Create a new LaTeX document from parameters.
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path for the new .tex file |
document_type | article|report|book|letter|beamer|minimal | article | Document class |
title | str | "" | Document title |
author | str | "" | Document author |
date | str | \today | Document date |
content | str | "" | Body content |
packages | list[str] | [] | Extra LaTeX packages |
geometry | str | "" | Geometry settings (e.g. margin=1in) |
create_from_template
Create a document from a bundled template.
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path for the new .tex file |
template | article|beamer|report | article | Template name |
edit_latex_file
Edit an existing LaTeX file.
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path to the file |
operation | replace|insert_before|insert_after|append|prepend | required | Edit operation |
new_text | str | required | Text to insert or replace with |
search_text | str|null | null | Text to find (required for replace/insert) |
line_number | int|null | null | 1-based line number (alternative to search_text) |
read_latex_file
Read and return the contents of a .tex file.
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path to the file |
list_latex_files
List all .tex files in a directory.
| Parameter | Type | Default | Description |
|---|---|---|---|
directory_path | str | . | Directory to search |
recursive | bool | false | Search subdirectories |
validate_latex
Check LaTeX syntax: required declarations, balanced braces, environment matching, undefined references.
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path to the file |
get_latex_structure
Extract document structure: class, title, author, packages, and section hierarchy.
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path to the file |
compile_latex
Compile a .tex file to PDF (runs the engine twice for references/TOC).
| Parameter | Type | Default | Description |
|---|---|---|---|
file_path | str | required | Path to the file |
engine | pdflatex|xelatex|lualatex | pdflatex | LaTeX engine |
Resources
| URI | Description |
|---|---|
latex://templates | List available bundled templates |
latex://template/{name} | Get the content of a specific template |
Testing
Test with the MCP Inspector:
uv run mcp dev latex_server.py
Or run the included test suite:
python test_server.py
Troubleshooting
Server won't start — Verify Python 3.10+ (python --version) and that mcp is installed (pip list | grep mcp).
Compilation fails — Ensure a LaTeX distribution is on your PATH (pdflatex --version). Install MiKTeX or TeX Live if missing.
"Access denied" errors — The requested file path resolves outside LATEX_SERVER_BASE_PATH. Use relative paths or adjust the env variable.
Claude can't connect — Double-check file paths in your MCP config, restart Claude Desktop, and verify with uv run mcp dev latex_server.py.
License
MIT
関連サーバー
Alpha Vantage MCP Server
スポンサーAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
MCP Agentic AI Crash Course with Python
A comprehensive crash course on the Model Context Protocol (MCP), covering everything from basic concepts to building production-ready MCP servers and clients in Python.
Remote MCP Server (Authless)
An example of a remote MCP server without authentication, deployable on Cloudflare Workers.
Squidler.io
Squidler is designed to validate your web app as a human based on natural language use cases, without write brittle, DOM-dependent tests.
MCP Jupyter Complete
A server for Jupyter notebook manipulation with position-based operations and VS Code integration.
Geo Location Demo
Retrieves user geolocation information using EdgeOne Pages Functions and integrates it with large language models via MCP.
SkillsMP
Search, discover, and install AI coding skills from SkillsMP marketplace with semantic search
SonarQube MCP Server
Integrates with SonarQube to provide AI assistants with access to code quality metrics, issues, and analysis results.
Code Assistant
A Rust-based CLI tool for code-related tasks, operating as an MCP server.
Prefect
Interact with the Prefect API for workflow orchestration and management.
Grafana
Search dashboards, investigate incidents and query datasources in your Grafana instance