kalyan9639/markdown-parser-4-ai

When you use an AI system, the response often comes back in Markdown format: plain text mixed with special characters that represent headings, bold text, code blocks, and lists. Raw Markdown looks fine in a chat interface but is messy if you want to save the output as a Word document, embed it in a web page, or pass it to another application as structured data. This Python package converts AI-generated Markdown into those three output formats. For HTML output, the package generates a two-column web page with a sticky table of contents on the left and the content on the right. It includes light and dark theme switching, automatic code syntax highlighting, copy-to-clipboard buttons on code blocks, support for mathematical notation, and diagram code rendered as vector images. The HTML it produces is self-contained and styled rather than plain converted text. For Word documents, the converter applies a corporate color scheme to headings, handles nested tables and lists without breaking the formatting, renders code blocks with color syntax highlighting inside the document, and converts checkboxes into the corresponding check symbols. Hyperlinks remain clickable in the resulting file. The styling can be overridden by passing a custom dictionary of font sizes, colors, and alignment settings. The JSON output is not a simple text dump. It walks the document structure and produces a block array that tracks each element's type and its character-level formatting, for example which portions of a sentence are bold, italic, or inline code. It also adds metadata including word count, estimated reading time, and a list of all links and code block languages found in the document. Installation is a single pip command pointing at the GitHub repository. The package exposes one entry point class that returns each format through separate method calls.