Choosing a File Format in DEVONthink
December 9, 2025 — Jim Neumann

In some discussions, DEVONthink is mentioned as a note-taking app. While this is not entirely accurate, you can indeed compose documents directly in DEVONthink in several file formats. The choice of format depends on your needs and, especially, your personal preferences. Here is an overview of the different options.
Open the Data > New menu to create a new document. You’ll see many options, but here are the available composition formats:
- Plain text: The simplest format, this is just text with no formatting. It is often considered the most future-proof format and is used for many purposes, e.g., quick drafts, logging, etc.
- Rich text: Rich text (RTF) is intended to create formatted documents without needing heavier word processing applications. It supports tables, lists, and attachments. It is used in many ways, from meeting minutes to journaling and many more.
- Formatted note: While not as feature-complete as rich text, this is an editable, HTML-based rich text format that can be viewed in a web browser. It supports basic formatting, e.g., bold and italics, but also images and adding checkboxes via Edit > Insert.
- HTML: These documents function like formatted notes, as you can directly edit the content. If you’re so inclined, you can also switch to the source code and hand-craft the underlying HTML.
- Markdown: Markdown is plain text to which you can add simple formatting markup. The source text renders as HTML and was designed as an efficient way to create web content while minimizing hand-coding. It is growing in popularity and is used for many of the same purposes as the previous formats. In a previous post, we already discussed why to use Markdown.
- Sheet: Sheets contain tabular data, presented as a table or as individual records in the Pro or Server editions of DEVONthink. With user-definable columns and data types, you can use them in many creative ways, like storing bibliographic information, genealogy records, etc. The format is intended for structured data and is less suitable for continuous text.