AI Interpretation & Translation Glossary
50 key terms related to AI interpretation, translation, and SEO explained in alphabetical order.
50 terms
A
- AI InterpretationAI 통역
- Technology where AI recognizes voice in real time and translates it into another language. Enables instant language conversion without a human interpreter.
- AI TranslationAI 번역
- Technology where artificial intelligence automatically converts text or documents into another language. InterLive uses a GPT-4o-based translation engine.
- ASR (Automatic Speech Recognition)자동 음성 인식
- Converts audio from a microphone into text. InterLive uses OpenAI Whisper for ASR.
- AEO (Answer Engine Optimization)AEO
- Applying FAQPage and HowTo structured data so content is directly cited by AI answer engines such as Google Featured Snippets.
- Automatic Language Detection자동 언어 감지
- A feature that automatically identifies the language being spoken. Enables multilingual conversation without manual switching in Face-to-face and Live modes.
- API (Application Programming Interface)API
- An interface for exchanging data and functions between software. InterLive uses the OpenAI API for speech recognition and translation.
B
- BreadcrumbListBreadcrumbList
- A schema that expresses a page's hierarchy path as structured data so it appears below the URL in search results.
C
- Conversation Mode대화 통역
- A push-to-talk two-way interpretation mode where two people take turns speaking by holding a button.
- Caption Viewer자막 뷰어
- A public link page provided by the meeting interpretation bot. Participants see real-time captions without logging in by clicking the link.
- Cloud RunCloud Run
- Google's serverless container execution environment. InterLive's meeting interpretation bot runs on Cloud Run.
- Chunk청크
- A segment of the audio stream divided into processing units. Audio data cut to a length suitable for Whisper recognition.
- Context Window컨텍스트 윈도우
- The length of text an LLM can process at once. Used in interpretation to remember previous utterances and improve translation consistency.
D
- DOCX TranslationDOCX 번역
- A feature that translates Microsoft Word (.docx) files while preserving paragraph, table, and footnote structure.
F
- Face-to-face Mode마주통역
- A two-way real-time interpretation mode where two people place one device between them and each speak their own language simultaneously. Consumes 2 tokens per minute.
- FAQPage SchemaFAQPage 스키마
- A schema.org structured data type that marks up FAQ content so rich snippets appear in Google search results.
- Footnote각주
- A format that appends additional explanation at the bottom of a document. Footnote text is also translated in InterLive DOCX and HWPX translation.
- Firebase파이어베이스
- Google's app development platform. Used in InterLive for user authentication and real-time database.
G
- GPT-4oGPT-4o
- A multimodal large language model from OpenAI. Used in InterLive for high-quality translation (Precise mode).
- GPT-4o-miniGPT-4o-mini
- A lightweight version of GPT-4o. Used in InterLive for Fast and Standard translation modes, offering high speed and cost efficiency.
- GEO (Generative Engine Optimization)GEO
- Optimising content to be cited by generative AI search engines such as ChatGPT and Perplexity.
H
- HWPX TranslationHWPX 번역
- A feature that translates Hangul Office HWP/HWPX files. Supports the format commonly used in Korean official documents and court filings.
- hreflanghreflang
- An HTML attribute that informs search engines of multilingual versions of the same page. An international SEO standard.
I
- Inference Level추론 강도
- A setting that controls AI translation quality and processing speed. Three levels: Fast (draft), Standard (default), and Precise (high quality).
- Interpretation Bot통역 봇
- Software that automatically joins a video conference and performs interpretation. InterLive's meeting bot is a prime example.
J
- JSON-LDJSON-LD
- A format for expressing Linked Data as JSON. The recommended format for Google Rich Results structured data.
L
- Live Interpretation대면 통역
- A one-directional interpretation mode where AI translates the speaker's voice in real time and outputs it as subtitles or audio. Suited for lectures and presentations.
- Layout-preserving Translation레이아웃 보존 번역
- A method of translating PDF, DOCX, and HWPX files while retaining formatting such as tables, paragraphs, and headers.
M
- Meeting Interpretation회의 통역
- A service where an AI bot automatically joins a Zoom, Teams, Meet, or Webex video conference and provides real-time interpretation captions.
- Multilingual Support다국어 지원
- The ability to process or switch between multiple languages. InterLive supports 15 languages.
- Monthly Free Interpretation월 무료 통역
- Free interpretation time automatically granted every month after signing up for InterLive. Currently 20 minutes per month.
- My Page마이페이지
- The user account page in InterLive. Displays session history, translated files, token usage, and billing records.
- Multi-column Layout다단 레이아웃
- A document layout with two or more columns. Common in Hangul files and difficult to perfectly preserve after translation.
- Middleware미들웨어
- A software layer that handles common processing between requests and responses. The InterLive Next.js middleware handles authentication and locale processing.
N
- NMT (Neural Machine Translation)신경망 기계 번역
- Machine translation based on neural networks. GPT-class models are a prime example, providing context-aware, high-quality translation.
O
- OCR (Optical Character Recognition)OCR
- Technology that extracts text from scanned images or image-based PDFs.
P
- Push-to-talk (PTT)푸시투톡
- A method where voice is transmitted only while a button is held. Adopted in InterLive's Conversation mode.
- PDF TranslationPDF 번역
- Technology that extracts text from a PDF file, translates it with AI, and reconstructs the original layout.
- Pay-as-you-goPay-as-you-go
- A billing model where you pay only for what you use. In InterLive, this corresponds to topping up and spending tokens.
R
- Real-time Streaming실시간 스트리밍
- A method of continuously transmitting data without interruption. A core technology for displaying interpretation captions without delay in InterLive.
S
- SEO (Search Engine Optimization)SEO
- The practice of improving content and structure so a web page ranks higher in search results.
- Structured Data구조화 데이터
- Data marked up in schema.org JSON-LD format so search engines can understand page content.
- Serverless서버리스
- A cloud computing model where code is executed without managing servers. Costs are incurred only for what is used.
- Session세션
- The unit of an interpretation run in InterLive, from start to end. The session history shows the original and translated transcript.
T
- TTS (Text-to-Speech)TTS
- Technology that synthesizes text into AI voice. Used to output translated results as real-time audio.
- Token토큰
- The usage unit for InterLive. Interpretation consumes 1–2 tokens per minute; translation is charged by character count.
- Token Exhaustion토큰 소진
- The state when all topped-up tokens have been used. Interpretation and translation stop and additional top-up is required from the dashboard.
- Transcript스크립트
- A record document pairing original speech with its translation from an interpretation session. Automatically generated after a session ends.
W
- WebSocketWebSocket
- A bidirectional real-time communication protocol between server and client. Used in InterLive to deliver interpretation subtitles without delay.
- WhisperWhisper
- An open-source speech recognition model developed by OpenAI. Recognises 15+ languages with high accuracy and supports automatic language detection.
- Web Audio APIWeb Audio API
- A web standard API for processing audio in the browser. Used by InterLive to capture microphone input.