Markdown Utilities Engine MCP. Stop formatting errors. Start structured content.

Listen, when you're dealing with complex documentation—the kind that needs perfect structure like a massive README or a technical spec—you know the drill. Just asking your AI client to spit out nested tables or a proper table of contents is risky business. General-purpose models struggle with maintaining structural integrity over long passages; they often mess up pipes, misalign columns, or generate broken link slugs.

This engine fixes that by giving your agent dedicated tools. You don't have to rely on the model just guessing how markdown should look; you use a specialized engine that guarantees clean output every time.

When you need to turn raw data into something readable—like turning an array of structured records into a slick, professional-looking table—you run generate_table_from_json. You feed it a standard JSON array containing objects. This tool doesn't just approximate the structure; it converts that entire JSON payload into perfectly formatted Markdown tables.

It automatically identifies all the required headers and correctly builds the row separators and column dividers, so you never gotta worry about broken pipes (|) or data getting misaligned across multiple rows. Think of it like having a specialized formatter: give it structured input, and it spits out rock-solid markdown that looks good right out of the gate.

If your documentation is huge—the kind where scrolling through 50 pages feels impossible—you need a Table of Contents, and you need it to actually work. That’s where generate_toc comes in. You drop in massive blocks of raw Markdown text, and this tool scans every single character. It detects all the header levels, whether they're H1 or H3, and then generates a complete, nested Table of Contents (TOC).

Crucially, it doesn't just list titles; it creates mathematically accurate link slugs for every detected header. This means when you click an entry in that generated TOC, your agent knows exactly where to jump in the document, giving you proper GitHub-style navigation right out of the box. The tool handles the parsing and linking simultaneously, meaning you get a cohesive index pointing directly to every section.

Using these tools lets you separate data structuring from language processing. Your AI client uses generate_table_from_json purely for formatting utility: take JSON, give clean markdown table output. It runs independently of general chat logic, making the result predictable and reliable. Similarly, when you need a navigable index, running generate_toc ensures that all those links are correctly slugged and nested, which is something models often botch up in simple conversational outputs.

You're not just formatting; you’re enforcing strict structural standards across your entire document set. It gives the agent highly reliable methods for converting structured JSON data into clean Markdown tables, and it also provides a mechanism to scan raw text blocks and generate nested lists of links that point precisely to every header slug detected.

This separation means your workflow is faster, cleaner, and much less prone to formatting errors you'd normally spend hours fixing with manual cleanup.