Markdown rendering too strict and incompatible with LLM-generated output

[jlzhjp]

Describe the bug

Hi, thanks a lot for creating this fantastic project — Marimo is really impressive and exciting to use!

I noticed that Marimo currently uses Python-Markdown to render markdown. However, the parsing rules seem to be a bit too strict and not always compatible with markdown generated by LLMs.

For example:

• A blank line is required before starting a list.
• Multi-level lists must use a four-space indent to render correctly.

This makes it harder to directly render markdown from LLM outputs, which usually follow more permissive conventions.

Would it be possible to consider adopting a more relaxed and widely supported standard, such as CommonMark or GitHub Flavored Markdown (GFM), for rendering? This could make LLM-generated markdown more compatible out of the box.

Thanks again for the amazing work!

Will you submit a PR?

• Yes

Environment

{
  "marimo": "0.16.1",
  "editable": false,
  "location": "/home/akari/anaconda3/envs/agent/lib/python3.12/site-packages/marimo",
  "OS": "Linux",
  "OS Version": "6.16.7-200.fc42.x86_64",
  "Processor": "",
  "Python Version": "3.12.11",
  "Locale": "en_US",
  "Binaries": {
    "Browser": "140.0.7339.185",
    "Node": "v22.19.0"
  },
  "Dependencies": {
    "click": "8.3.0",
    "docutils": "0.22.2",
    "itsdangerous": "2.2.0",
    "jedi": "0.19.2",
    "markdown": "3.9",
    "narwhals": "2.5.0",
    "packaging": "25.0",
    "psutil": "7.1.0",
    "pygments": "2.19.2",
    "pymdown-extensions": "10.16.1",
    "pyyaml": "6.0.2",
    "starlette": "0.48.0",
    "tomlkit": "0.13.3",
    "typing-extensions": "4.15.0",
    "uvicorn": "0.35.0",
    "websockets": "15.0.1"
  },
  "Optional Dependencies": {
    "nbformat": "5.10.4",
    "openai": "1.108.1",
    "loro": "1.8.1",
    "python-lsp-server": "1.13.1",
    "ruff": "0.13.1"
  },
  "Experimental Flags": {}
}

Code to reproduce

import marimo

__generated_with = "0.16.1"
app = marimo.App(width="medium")


@app.cell
def _():
    import marimo as mo
    from langchain_openai import ChatOpenAI
    from dotenv import load_dotenv, find_dotenv

    load_dotenv(find_dotenv(usecwd=True))

    llm = ChatOpenAI(model="gpt-4.1")
    response = llm.invoke(
        [
            ("system", "You are a helpful assistant"),
            ("human", "Use a brief paragraph and a multilevel unordered list to show the advantages of marimo over Jupyter."),
        ]
    )
    return mo, response


@app.cell
def _(mo, response):
    mo.md(response.content)
    return


@app.cell
def _(response):
    print(response.content)
    return


if __name__ == "__main__":
    app.run()

[mscolnick]

This is a good callout - we can definitely add support for more GFM. I'll make sure it makes it in the next release