Glossary: 104 terms about documents and AI

A

Archived copy

A saved version of a web page at a specific point in time, for example in the Wayback Machine. → Web pages as sources

B

Backup

An additional copy in another place, for example via Time Machine or Windows Backup. → Local drafts

Balance

The credits still available to you in the current billing period. → Credits and devices

Bar chart

Compares values through the length of bars; the axis should start at zero. → Charts

Base

The total a percentage refers to. → Analysing spreadsheets

Billing period

The month Stripe bills your subscription for and for which your allowance applies. → Credits and devices

Bookmarks

The outline of a PDF in the sidebar. Docsi exports create them from the headings. → Merging PDFs

C

Character encoding

The rule for storing characters as bytes, such as UTF-8. → Checking CSV data

Checksum

A fingerprint of a file (SHA-256 in Docsi). If the file changes, the checksum changes. → Reviewing sources

Citation style

A set format for references, such as APA or Harvard. Docsi provides file and page; you add author and year. → Reviewing sources

Context

All content an AI task sees: the message, earlier steps, source excerpts and remembered details. → AI and privacy

Contrast

The brightness difference between text and background; high contrast improves recognition. → Text from images

Conversation

A connected sequence of requests. Docsi considers up to 16 earlier steps in it, but no other conversations. → Getting started

Credit

The unit of usage for the built-in Docsi AI. Local tools use no credits. → Getting started

Cropping

Removing unnecessary image margins so the relevant text stays larger. → Text from images

CSV

A text file holding one sheet, with fields separated by comma, semicolon or tab. → Converting files

D

Data minimisation

A GDPR principle: process only as much personal data as necessary. → AI and privacy

Data series

A sequence of related values, such as one year’s revenue per month. → Charts

Date accessed

The date you read a web page; important for pages that change. → Web pages as sources

Deadline

The date by which a task should be completed. → Meeting notes

Decision

An explicit agreement of the group, recorded in the minutes. → Meeting notes

Delimiter

The character between the fields of a CSV line: comma, semicolon or tab. → Checking CSV data

Dependency

A passage that relies on another, such as a definition later sections refer to. → Comparing versions

Device seat

An activated computer. Plus has 2 seats, Pro 3. → Credits and devices

Digital signature

A cryptographic proof for an unchanged file. It does not apply to a newly merged PDF. → Merging PDFs

Direct quote

A sentence taken over unchanged in quotation marks, with a page number in the citation style. → Reviewing sources

Docsi AI

The built-in AI based on the GLM language model by Z.ai, reached through the Docsi server. → AI and privacy

DOCX

The Word format following the open Office Open XML standard (ECMA-376). → Converting files

Donut chart

Shows parts of a whole as a ring; suitable only for a few categories. → Charts

dpi

Dots per inch, a measure of scan resolution. Around 300 dpi is a good guide for normal text. → Text recognition

Draft

An intermediate state of a result in the app’s local working store. → Local drafts

E

Embedded font

A typeface contained in the file so the document looks the same everywhere. → Export formats

Evidence

The passage in a source that a statement relies on. → Reviewing sources

Export

A file you saved in PDF, Word, Excel, CSV or Markdown format. → Local drafts

F

Field

A single value in a CSV line, equivalent to a cell. → Checking CSV data

File signature

The first bytes of a file, which reveal its actual type regardless of the extension. → Protected files

Flashcard

A pair of question or term and answer. Docsi lays cards out in pairs in the PDF and delivers them as a table on request. → Study sheets

Formula cell

A cell containing a formula. Docsi does not count it by default because its stored result may be outdated. → Analysing spreadsheets

G

Grouping

Combining rows by a feature, such as the sum per region. → Analysing spreadsheets

H

Header row

The first line of a CSV file with the column names. → Checking CSV data

HEIC

The default photo format of many iPhones; convert to JPEG before adding. → Text from images

I

Interruption

A task ends before the result is complete, for example due to a lost connection or cancelling. → Interrupted tasks

ISO 4217 code

The three-letter currency code, such as EUR, USD or CHF. → Receipts into tables

ISO date

A date in year-month-day format, such as 2026-10-29, unambiguous in every language. → Checking CSV data

K

Keychain

The operating system’s protected credential store, where Docsi keeps keys. → AI and privacy

L

License key

The key in the format DOCSI-XXXXX-XXXXX-XXXXX-XXXXX that links a device to your subscription. → Credits and devices

Line chart

Shows a trend along an ordered axis, usually time. → Charts

Local model

A language model running on your computer, for example through Ollama or LM Studio. → AI and privacy

Local model server

A program such as Ollama or LM Studio that serves a language model on your computer. → Interrupted tasks

LSTM

The neural recognition method of Tesseract that Docsi uses exclusively. → Text recognition

M

Macro

A program embedded in Office files. Docsi rejects files with macros. → Protected files

Markdown

Plain readable text with characters for headings, lists and tables, such as “#” and “|”. → Converting files

Mean

The sum of all values divided by their count. Sensitive to outliers. → Analysing spreadsheets

Median

The middle value of a sorted series. Robust against outliers. → Analysing spreadsheets

Merging

Copying the pages of several PDFs unchanged one after another into a new file. → Merging PDFs

Metadata

Information about a file such as title, language and date that is not part of the visible text. → Export formats

Modal word

Words such as “must”, “should” or “may” that set the degree of an obligation. → Comparing versions

N

Native charts

Charts that Word or Excel recognise as their own objects and can edit. → Export formats

Net and gross

Net is the amount without VAT, gross the amount including VAT. → Receipts into tables

O

OCR

Optical character recognition: recognising text in images and scanned pages. → Text recognition

Open password

Encrypts a file; without the password it cannot be read. → Protected files

Original

The original file, which Docsi only reads and never changes. → Local drafts

Owner

The person who completes a task or makes sure it gets done. → Meeting notes

P

Paraphrase

A statement from the source in your own words. It also needs a source reference. → Reviewing sources

Partial draft

An incomplete result that can be reused as a working state. → Interrupted tasks

PDF/A

A standardised PDF variant for long-term archiving under ISO 19005. → Merging PDFs

PDF/UA

A standard for accessible PDFs. Docsi does not claim PDF/UA conformance. → Export formats

Permissions password

Allows a PDF to be opened but restricts printing, copying or editing. → Protected files

Proration

On a plan change, Stripe charges only the difference for the remaining days. → Credits and devices

R

Reader view

A browser view that shows only the article text and hides ads. → Web pages as sources

Receipt

Proof of a business transaction, such as an invoice, receipt or till slip. → Receipts into tables

Recognition confidence

An engine estimate between 0 and 100. It is not a probability that the text is correct. → Text recognition

Recorded changes

A Word feature that records changes visibly until they are accepted or rejected. → Comparing versions

Retrieval practice

Learning by actively recalling, for example with review questions or flashcards, instead of re-reading. → Study sheets

Review question

A question that checks whether you understood a section, ideally with an answer and page reference. → Study sheets

S

Screen recording permission

The macOS permission an app needs to capture other windows. → Windows as sources

ScreenCaptureKit

Apple’s interface for screen and window captures on macOS. → Windows as sources

Screenshot

An image of the screen content or part of it. → Text from images

Small invoice

An invoice up to 250 euros, for which simplified required details apply. → Receipts into tables

Source

A file, window or pasted text Docsi reads for a task. Sources are numbered (Q1, Q2 …). → Getting started

Source marker

A reference in the result to a passage of a source, with the page number for PDFs. Docsi checks every marker against the original. → Getting started

Spaced repetition

Reviewing at intervals over several days instead of in one long block. → Study sheets

Statement title

A chart title that states the key message instead of just the topic. → Charts

Status code 429

The HTTP response “Too Many Requests”: a service is not accepting further requests right now. → Interrupted tasks

Storage limitation

GDPR principle: keep personal data only as long as the purpose requires. → Local drafts

Study sheet

A short, structured summary of a topic with key terms, takeaways and questions for practice. → Study sheets

Summarising

Shortening the content with AI. That is a different workflow from merging. → Merging PDFs

T

Task

A concrete action with an owner and a date, often called an action item. → Meeting notes

Technical diff

A list of all text differences between two files, for example with “Compare” in Word. → Comparing versions

Text layer

The invisible, selectable text of a PDF. If it is missing, the page consists only of an image. → Text recognition

The selectable text of a PDF; a PDF saved from the browser usually has one. → Web pages as sources

Timeout

A service does not respond within the expected time; the connection is closed. → Interrupted tasks

Transcript

A verbatim written version of what was said. Docsi does not create transcripts but can work with existing ones. → Meeting notes

Typesetting

Arranging text, tables and figures on pages. Docsi sets PDFs with its own layout engine. → Export formats

U

URL

The address of a web page, such as example.com/article. → Web pages as sources

UTF-8

The most widespread character encoding, covering accented letters and all Unicode characters. → Converting files

V

VAT

The tax on goods and services, in Germany 19 percent standard and 7 percent reduced. → Receipts into tables

Version

A particular state of a document, ideally named with date and role. → Comparing versions

Visible area

The part of a window currently shown on screen. Only this part is captured. → Windows as sources

Vision model

An AI model that can process images as well as text. → Text from images

W

Widget

The slim Docsi bar at the top of the screen. It opens when you hover over it and accepts files, windows and quick questions. → Getting started

Window drop

Dragging a window onto Docsi so he takes an image of it and reads the text. → Windows as sources

Windows.Graphics.Capture

The public Windows interface for screen and window captures. → Windows as sources

Working copy

A separate copy for further work, so the original stays unchanged. → Protected files

X

XLSX

The Excel format under Office Open XML with several sheets and typed cells. → Converting files

How this glossary was made

The explanations come from the “Key terms explained” sections of the Docsi AI guides. They describe the terms as they come up when you work with Docsi and do not replace a technical definition. The primary sources for formats, standards and methods are listed at the end of each guide.