openclaw - 💡(How to fix) Fix [Feature]: Latex Rendering in the WebUI chat

Summary

Render LaTeX math expressions as formatted mathematical notation in the Web UI instead of displaying raw dollar-delimited text.

Problem to solve

Currently, inline and block LaTeX expressions are displayed as raw source text (for example $E = mc^2$ or $$\int_a^b f(x),dx$$) instead of rendered mathematical notation.

This becomes a significant readability issue when using science- and math-oriented models that frequently generate mathematical expressions. In my case, models such as Gemma 4 31B rely heavily on LaTeX-style notation in normal conversation and explanations. Raw LaTeX output makes longer technical chats noticeably harder to read, parse, and learn from.

The current behavior is especially limiting for:

Mathematics discussions Physics and engineering explanations Chemistry notation Educational use cases Research-oriented workflows

At the moment, users either need to mentally parse raw LaTeX or manually copy formulas into external renderers, which interrupts workflow and reduces usability.

Proposed solution

Add native LaTeX math rendering support to the Web UI markdown pipeline.

Desired behavior

Support both:

Inline math using:

$...$

Block math using:

$$...$$ Example

Input:

The quadratic formula is $x=\frac{-b\pm\sqrt{b^2-4ac}}{2a}$

$$ \int_0^\infty e^{-x} dx = 1 $$

Expected rendering:

Inline math rendered directly within text flow Block equations rendered centered and visually separated

Alternatives considered

i think its unavoidable

Impact

Affected users/systems/channels

Affected users include:

Users running local STEM-oriented models Technical/research users Students and educators Anyone discussing math, physics, chemistry, statistics, or engineering topics

Primary affected channel:

Web UI chat interface Severity

Moderate to high.

This does not fully block usage, but it substantially reduces readability and overall UX quality for technical conversations.

For math-heavy workflows, it becomes a constant source of friction.

Frequency

Very frequent for technical models and STEM discussions.

Some models output LaTeX in nearly every response.

Consequences

Current consequences include:

Reduced readability Slower comprehension Manual mental parsing of formulas Need for external rendering tools Increased cognitive load during learning or research Less polished user experience compared to platforms with native math rendering

Evidence/examples

No response

Additional information

No response