Formatting

How markdown in your reply becomes Zoom content segments — including what Zoom cannot carry and what is dropped on the way.

Zoom Team Chat does not take markdown. A message is a list of segments, each one a string with an optional bold or italic style, so the adapter converts your reply into that shape before sending it.

Most of what you write survives. Some of it does not, and the table below is the honest version.

What Zoom receives

Zoom shows

Deploy finished

Shipped `api-gateway` to production in 94s.

What reaches Zoom

What you write	What Zoom receives
A paragraph	One segment of plain text
A paragraph that is bold end to end	One segment with `style.bold`
`bold` inside a sentence	The words, with the markers removed and no bold
`_italic_`	The words, with the markers removed and no italics
`code`	Kept as written, backticks and all
`# Heading`	A bold segment
`* item`	A segment starting with `•`
A fenced code block	One segment, fences included
`~strikethrough~`	Dropped — the text does not arrive
`__underline__`	Dropped — the text does not arrive
`[label](https://…)`	Dropped — neither label nor URL arrives
`:tada:` and other shortcodes	The Unicode emoji
An empty message	Nothing is sent; the Zoom API rejects empty messages

Inline bold and italics are read and then flattened: **Deploy** finished arrives as Deploy finished with no emphasis. To get a bold line in Zoom, make the whole paragraph bold:

await thread.post("**Deploy finished**\n\nShipped api-gateway in 94s.");

That sends two segments — a bold one and a plain one.

Reading an incoming message

The other direction is lossless. Incoming text is parsed with Zoom’s own conventions, which differ from standard markdown in two places:

Token	Meaning
`__text__`	Underline, not bold
`~text~`	Strikethrough, with a single tilde

Everything else — **bold**, _italic_, `code`, # headings, * lists — is standard. message.text holds the raw text the user typed, and message.formatted holds the parsed tree.