For Developers

Prompt Template

Fix LaTeX dollars

Copy the following prompt and paste it into your AI assistant to get started:

AI Prompt

Investigate and fix the actual $ usages in Markdown content.

The $ fall into three classes:

- Currency (escape these) — $1, $2 billion, R$ 549 → these pairs cause all the warnings
- Real math (leave alone) — $\rightarrow$, $O(1)\text{ streaming}$ → valid, no warnings
- Shell code (leave alone) — $(curl…), ${ZSH_CUSTOM}, $HOME → inside code blocks


Execute in 4 steps:

- Investigate — greps the content, classifies every $ into currency / real math / shell code, and reports counts before changing anything.
- Apply — checks the tree is clean, then writes and runs the exact tested Python script (code-fence-, inline-code-, frontmatter-, and math-span-aware; idempotent via the (?<!\\) lookbehind so re-running never double-escapes).
- Verify the diff — the safety net: greps that must print nothing for real math ($\rightarrow$, \text) and shell vars ($HOME, $(…), ${VAR}). If anything legit was touched, it tells you to git checkout -- . and stops.
- Print instructions — outputs the build-verify and commit/push commands for user to run.

Do not autonomously run any build, commit, or push.
Try Prompt

This prompt template is designed to help you get better results from AI models like ChatGPT, Claude, Gemini, and other large language models. Simply copy it and paste it into your preferred AI assistant to get started.

Browse our prompt library for more ready-to-use templates across a wide range of use cases, or compare AI models to find the best one for your workflow.

AI chat subscription

Turn model research into daily AI work.

Use 40+ models, web search, files, and EU-hosted options in one paid chat workspace.

Inference credits

Build with EU-hosted open-source models.

OpenAI-compatible API for GLM, Kimi, DeepSeek and more. Add credits inside the dashboard.