typst

Typst MCP Server exposes tools that help AI models interact with Typst, a markup-based typesetting system. The server provides capabilities for exploring Typst documentation, converting between LaTeX and Typst, validating Typst syntax, and rendering Typst code as images. Key tools include: list_docs_chapters to enumerate Typst documentation chapters, get_docs_chapter(route) to fetch a specific chapter, latex_snippet_to_typst(latex_snippet) to convert LaTeX to Typst with Pandoc, check_if_snippet_is_valid_typst_syntax(typst_snippet) to validate Typst code, and typst_to_image(typst_snippet) to render Typst code to a PNG. These tools are designed to be used by LLMs to understand and generate Typst content more effectively, with optional batch variants like get_docs_chapters(routes) and latex_snippets_to_typst(latex_snippets) for convenience.

To use the server, start it via your preferred deployment method (Docker in most setups). Once running, you can query the available tools to retrieve Typst documentation, validate code, and generate visuals. When integrating with AI workflows, chain the tools to first pull relevant documentation chapters, convert or validate sample Typst code, and render previews as needed before presenting results to end users.

Prerequisites:

• Docker installed and running (recommended) or a local Python environment if you prefer running server.py directly.
• Network access to pull the Typst MCP Docker image or to run a local server.

Option A: Run with Docker (recommended)

1. Ensure Docker is installed and running on your machine.
2. Pull and run the Typst MCP server image:

docker run --rm -i ghcr.io/johannesbrandenburger/typst-mcp:latest

3. The server should start and listen for MCP requests from your integration.

Option B: Run locally with Python (server.py)

1. Clone the repository and install dependencies if needed.
2. Run the server locally:

python server.py

3. If you prefer a local MCP workflow in Claude/OSS integrations, you can register the local server using:

mcp install server.py

Notes:

• If you already have the Typst MCP image locally, you can modify the Docker command accordingly to use a specific tag or a local image name.
• Ensure your client configuration points to the correct command and arguments for your environment (as shown in the example configurations).

Tips and common issues:

• The Docker command uses --rm and -i to keep containers ephemeral and interactive, which suits MCP usage. If you need persistence or volume mounting for assets, adjust the docker run command accordingly.
• If you switch to a local Python server, ensure Python 3.x is installed and that server.py is executable in your environment.
• When using various platform integrations (Claude Desktop, Cursor, Roo Code, OpenCode, Claude Code, VS Code Agent Mode), replicate the docker-based configuration or the local python execution as appropriate for the platform instructions.
• If you encounter network or permission issues, verify that the Docker daemon has proper access and that your firewall allows container traffic.
• The image ghcr.io/johannesbrandenburger/typst-mcp:latest is the recommended entry point and is updated to reflect Typst MCP capabilities.