What AI Agents does with a Markdown file, and where a legacy Word document fits

AI Agents takes Markdown through a file sitting in the working directory the agent is allowed to read, and that is the door this page is about. The question worth asking is what a legacy Word document looks like by the time it arrives, because AI Agents never sees the document itself. It sees whatever the conversion handed over, and nothing else.

AI Agents reads a heading tree as a map of where to look next, and skips whatever the map does not mention. What comes out of a legacy Word document here is headings, bold, italics, nested lists and tables, in the order the document had them, so the structure AI Agents is looking for is actually present instead of merely implied. That is the entire trick, and there is nothing clever underneath it.

What survives from a legacy Word document into AI Agents

The conversion drops field codes, revision marks, print settings and the styles nobody ever applied before AI Agents ever sees them, because none of it carries meaning into AI Agents and all of it costs something. What it keeps is the part you would otherwise have retyped by hand.

• Heading levels and list nesting survive, which is what makes the document navigable afterwards, and AI Agents receives it as structure rather than as something to infer.
• Bold and italics survive, so emphasis that carried meaning still carries it, which is the part AI Agents would otherwise have to reconstruct on its own.
• Tables survive as tables rather than as runs of tab separated text, so nothing downstream of AI Agents has to guess at it.

What AI Agents never gets back from a legacy Word document

Every conversion costs something, and a page listing only the gains is a sales page. Here is what a legacy Word document gives up on the way to AI Agents, stated plainly enough that you can decide against it.

• Tracked changes and comments do not come through, and they are often the interesting part, and AI Agents will not get it back from anywhere.
• Page setup and fonts are gone, along with anything that depended on where a page ended, so nothing you build in AI Agents should depend on it.
• Objects embedded rather than written, such as a spreadsheet dropped into the page, are not unpacked, which is a real cost inside AI Agents and not a rounding error.

The Word to Markdown step behind this AI Agents page was measured

LibreOffice 24.2 does the conversion, and it was proven end to end on 13 August 2026: binary .doc, .rtf and .odt all came back with headings, bold, italics, lists and tables intact, and a 2 MB RTF of 4000 sections and 80 tables came back in 3.1 seconds. That is one run on one day and not an average, which is exactly why the date travels with the claim instead of being left out of it. The same output is what AI Agents would have received.

Above 25 MB the answer is a refusal, and a document LibreOffice cannot open is refused outright rather than returned empty. AI Agents has no way of telling you any of that from the inside, so it gets said here instead, on the page that sent you there.

Once the document is Markdown, what to do inside AI Agents

With the Markdown in hand, keep one Markdown file a subject rather than one long file, because an agent pays for everything it opens. That is a AI Agents habit rather than a conversion setting, and it is where most of the value of converting a legacy Word document in the first place actually lands.

The constraint worth knowing before you start: an agent re-reads the same file at every step it takes, so the size of that file is a recurring cost. It bites the same way whether the Markdown came out of a legacy Word document or out of something else entirely, but it bites hardest on long documents, which is what a legacy Word document usually is.

What we will not pretend about Word in AI Agents

• AI Agents will act on a mangled sentence rather than stop and ask about it, and converting a legacy Word document first does not change that in the slightest.
• a loop that reads a noisy file ten times pays for that noise ten times, which is a AI Agents behaviour rather than anything the document conversion did.
• The conversion costs half a credit and the price is shown before the document is sent, whether the Markdown ends up in AI Agents or nowhere at all.