Markdown in Comments and Descriptions
Dataset descriptions and object comments - set with COMMENT ON TABLE/COMMENT ON VIEW, or edited directly from the web UI's Manage Dataset page - support a limited subset of Markdown when rendered.
Supported syntax
- Bold -
**text** - Italic -
*text* Inline code-`text`- Links -
[text](https://example.com). Onlyhttp://andhttps://links are rendered; anything else is shown as plain text. - Headings -
#through#### - Bulleted lists - lines starting with
-or* - Numbered lists - lines starting with
1.,2., and so on - Tables - GFM-style pipe tables, with a
|---|---|header separator row - Fenced code blocks - triple backticks, with an optional language hint
Not supported
- Nested/indented lists
- Blockquotes
- Images
- Strikethrough
- Task lists
- Raw HTML
Unsupported syntax is not stripped - it's shown as literal text.
Example
Raw event records ingested from the upstream tracking pipeline.
- Deduplicated by `id`
- Backfilled from **2026-01-01** onward
- See the [ingestion runbook](https://example.com/runbook) for details
Notes
- Everything is escaped before formatting is applied, so comment/description text can never inject HTML or scripts into the page.
- Because comments are set through SQL string literals, a literal single quote in the text must be escaped by doubling it (
'') - this is standard SQL string escaping, not part of the Markdown syntax. See COMMENT.