Bars: Smaller sections of a window providing specialized information or controls, e.g., the Information Bar found just above the item listing. It shows information about the selection as well as buttons to sort or change the view.

Context menu: The menu that appears when you Control-click something. The context menu gives you access to frequently used commands that deal directly with the clicked object. Specific context menu items are discussed in their respective sections, but there is a list of commonly seen commands in the Context Menu Items section of this appendix.

Inspectors: Specialized panes at the right of a main window, providing information about the current selection as well as functions like See Also & Classify.

Item list: The list of items in the current location, viewable as a list, icons, columns. Items can be documents, groups, or smart groups.

Popover: A specialized type of window that appears at a specific location, e.g., the toolbar, opened via a toolbar button or menu command. Database Properties is a type of popover.

Sidebars: The panes found at the left side of a main window, comprised of the Navigate, Reading List, Import, Extras sidebars.

View: The layout of the item list and view/edit pane in a main window, either Standard or Widescreen. You can change the layout using the commands found in the View menu.

View/Edit pane: The pane in main windows where the content of selected documents is shown or edited. This pane can be hidden by choosing View > None.

Window: DEVONthink uses three different types of windows: main, document, and other. Main windows show lists of items and, in some views, also the contents of the selected document, whereas document windows can be opened for individual documents.

Database: A collection of all documents stored in DEVONthink. You can create and open as many databases as you like.

Document: A file in the database; for example, a plain or RTF text, a PDF, or an image (.jpg, .tiff, .png, etc.). Technically a document is an entry in a database with text, RTF, or image content, sometimes simply referred to as "content".

Default destination: The default destination group for incoming data. Set in Settings > Files > Import.

Duplicate: A copy of a file, either explicitly created or detected by DEVONthink's AI.

Group: A collection of documents or other groups inside the database. Similar in appearance and behavior to a folder in the Finder. Read more...

Global Inbox: A special database serving as a destination for incoming data. The Global Inbox is always opened when DEVONthink is started and accessible through the sidebar. The Global Inbox can be set as the default destination in Settings > Files > Import.

Inbox: A special group for data coming in to a database. Every database has an Inbox group.

News feed: Also referred to as RSS, a news feed is like a web page, but split into multiple articles. A feed reader or web browser can split the feed "page" into its parts and show each of them as a separate news item and/or store it in a local database. DEVONthink stores feeds in a database and displays them like a group. The contents are updated live from the internet as the connected feed changes. Older news bits are kept, while fresh news bits are added and marked as unread. When you add a bookmark that points a feed, DEVONthink tries to add it as a proper feed document if possible.

Replicant: A document or a group that appears in more than one place in the database. Unlike an alias in the file system, a replicant is not a real document pointing to another file (usually the original); rather it is a second entry in the table of contents for the very same document. Consequently, there is no original; when you replicate a document you'll end up with two replicants.

Sheet: A collection of records or structured sets of data, presented in a spreadsheet style, similar to data viewed in Apple Numbers or Microsoft Excel. Alternatively, the data can be displayed in form view, similar to browsing records in, e.g., FileMaker.

Smart group: A special type of group with a set of search criteria saved with it. This is similar to "Saved Search" folders in the Finder. Every time you visit this smart group, DEVONthink runs the attached search and displays it as the content of the group. DEVONthink creates a number of pre-defined smart groups when you create a new database.

Trash: A temporary location in a database for deleted files. Similar to the Trash in the Finder, all trashed items are stored until you explicitly delete them using DEVONthink > Empty Trash.

Group tag: Groups within a database that apply their name as tags to their contents. Items moved out of these groups will not retain the group tags. Group tags appear grey in all tag lists.

Ordinary tag: Tags created through the Tag bar or other means are created in a special top-level Tags group. They are called "ordinary tags" and appear blue in all tag lists.

Flagged/Unflagged: Items can be checked or unchecked for your personal use, e.g., for creating a to-do list, maintaining a shopping list, or writing a list of project tasks.

Locked/Unlocked: Items can be locked to keep them from being unintentionally deleted (groups, documents) or modified (documents).

Read/Unread: Items can be marked read or unread, which is useful for captured or automatically downloaded news items or documents that you want to read later. Both unread items and groups containing unread items are shown in bold.

UUID: Each database, group, or document in DEVONthink receives a unique identifier (UUID) when it's created. This is a read-only alphanumeric string that refers only to the particular item. The UUID allows you to reference an item regardless if its location or name changes, and is usually found in item links.

Sync location: Sync locations are used to keep your databases on all your devices in sync. They summarize all information about which databases to sync through what kind of connection (e.g., direct connections to DEVONthink on a Mac, WebDAV, CloudMe, Dropbox etc.) and with which sync store.

Bonjour (Direct connection): A sync method for syncing one device to another Mac or mobile device running DEVONthink or DEVONthink To Go, respectively. Unlike other sync methods, there is no intermediary storage location for the sync data. It is direct - device to device.

Sync store: A sync store is a special folder containing sync data. These can be direct, local, e.g., on your Mac or a thumb drive, or remote, as used with cloud services.

Encryption key: A user-defined word used to "scramble" or "unscramble" your sync data when syncing. When used, your sync data is stored in an encrypted state in the sync store.

Local database: A database that exists and is open on this device.

Remote database: A database that is not open on this device, or exists at the sync location, e.g., on Dropbox, but has not been downloaded to this device.

Pending document: A document for which only the metadata is locally available. This is usually caused by choosing not to synchronize indexed files for the sync location.

WebDAV: A type of web server used by some cloud services, some network attached storage devices, sometimes running locally. It can often be used by DEVONthink for syncing.

AI model: An umbrella term for different types of AI, e.g., text generation with ChatGPT, image generation via DALL-E.

API Key: This is a personal key provided by an AI provider, e.g., OpenAI, that lets you access and use their servers and data to chat with. You enter this into the AI > Chat settings.

Context window: The context window is how much information an LLM can process. This can include the content of previous queries. Consider it the short term memory of an LLM. The larger the context window, the more information it can use to generate appropriate responses.

Generative AI: A type of AI capable of creating images, videos, music, etc. DEVONthink supports image creation.

Hallucination: An incorrect or irrelevant AI response. This is less common as large commercial LLMs improve but still possible as the responses aren't validated. This is much more common when trying to run a local LLM with a small model.

LLM: The acronym for Large Language Model, focused on text processing and generation. ChatGPT and Claude are LLMs.

Prompt: This is the question or command you set to your AI chat, e.g., via the Chat inspector, Chat popover, or even AI assisted automations.

RAG: The acronym for Retrieval Augmented Generation, this is the process by which a document is parsed by an AI model into a specialized database. This database is then used to answer questions about the specific document. These are generally only available in commercial AI models, like Claude.

Role: A "persona" sent to the AI used to direct the responses, either on the part you or the AI. For example, your role could be "I am in an undergraduate anatomy class…" or "You are presenting to a roomful of young astronomers…".

Speech-to-Text: A type of AI capable of converting speech in audio and video files into text. DEVONthink supports this.

Tokens: Tokens are how words are processed by AI. Some words are treated as a token; others are split into separate tokens. This is controlled by how the LLM is built. Bear in mind, you typically pay per token when using a commercial AI service though the cost-per-token is usually small.

Tools: AI capable of accepting DEVONthink-specific commands, e.g., document creation, database searching, etc.

Reasoning: AI using a deeper, recursive approach to responding to a question, often returning a preamble of its "thinking".

Vision: AI capable of analyzing images, e.g., summarizing an image.