새로운 회의 에디터

이제 회의에 전용 워크스페이스가 제공됩니다. 편집 가능한 노트를 중심으로 노트, 타임라인, 전사 텍스트가 나란히 배치됩니다.

• 구조화된 노트：회의의 실제 내용에 맞춰 요약, 결정 사항, 액션 아이템, 미결 문제를 단락 형태로 제공합니다.

• 모먼트 추적：각 항목에는 근거 내용이 포함되며 녹음의 해당 시점과 연결됩니다. 클릭 한 번으로 당시의 발언을 다시 들을 수 있습니다.

• 타임라인 정렬：노트가 회의 타임라인을 따라 생성되어 2시간 분량의 대화를 완벽하게 커버합니다.

• 글로벌 화자 이름 지정：화자 이름을 한 번만 지정하면 전사 텍스트, 타임라인, 모든 노트에 적용되며 액션 아이템 담당자까지 동기화됩니다.

• 타임스탬프가 포함된 AI 어시스턴트：에디터의 어시스턴트가 녹음을 바탕으로 답변하며, 모든 설명에는 클릭 가능한 타임스탬프가 첨부됩니다.

• 통합 버전 기록：전사 텍스트, 번역 및 모든 노트를 복구 가능한 하나의 버전으로 기록합니다.

• 실시간 협업：탭, 기기 또는 공유 링크에 관계없이 동일한 실시간 문서에 액세스할 수 있으며, 편집 내용은 서로 덮어쓰지 않고 자동으로 병합됩니다.

더 편리해진 실시간 자막

• 간소화된 프롬프트 제어：용어집과 프롬프트 제어 항목이 기본값이 포함된 단일 프롬프트 필드로 통합되어 한 번에 설정을 마칠 수 있습니다.

• 모국어 명칭 메뉴：시청자 언어 선택 메뉴가 각 언어의 고유 명칭과 문자로 표시됩니다.

• 고급 재생 기능：이중 언어 자막과 실시간 번역 음성을 지원합니다(시스템에서 생성된 음성이 실제로 재생됩니다).

• PIP(Picture-in-Picture) 스트리밍：독립된 스트림으로 자막을 계속 표시합니다.

• 스마트 언어 처리：중국어 및 광둥어 타겟 언어는 문어체로 출력되며, 용어집 단어는 변환 후에도 유지됩니다.

마음대로 시작되지 않는 회의 녹화

• 수동 제어：이제 녹화와 실시간 자막 모두 수동으로 시작해야 합니다. 통화 참여 시 요청하지 않은 동작이 실행되지 않습니다.

• 강력한 녹화 도구：Zoom 회의 프레임 내 진입, 미디어 요소가 없는 페이지 캡처가 가능하며 캡처 실패 시 타이밍을 맞춰 복구합니다.

• 대기열 자동 복구：이전에는 봇이 시작되지 않았던 반복 회의가 이제 자동으로 대기열에 다시 추가됩니다.

막힘없는 자막 편집

기존에 고객 센터 문의가 필요했던 작업들을 이제 버튼 클릭 한 번으로 완료할 수 있습니다:

• 완료된 번역 트랙을 직접 삭제하거나 다시 생성할 수 있습니다.

• 잘못 들은 단어를 한 번만 수정하면 전체 트랙에 자동으로 적용됩니다.

• 자막 스타일 디자인 모드를 도입했으며, 낮은 화면에서도 접근 가능한 컴팩트 도구 상자를 제공합니다.

• 자막을 드래그할 때 인접한 자막의 가독성 길이를 유지하여 너무 짧게 쪼개지는 것을 방지합니다.

더 정확한 전사

• 정확도 대폭 향상：정확도가 한 단계 더 업그레이드되었으며, 기본적으로 전체 오디오를 사용하여 시간을 재정렬하므로 자막이 말소리에 정확히 밀착됩니다.

• 스마트 언어 프롬프트：자동 및 다국어 프로젝트에서 언어가 예측될 때 프롬프트를 제공합니다.

• 형식 정규화：중국어와 일본어의 공백이 정규화되고, 한국어는 실제 표시 너비로 계산되며, 최종 자막을 나누기 전에 불필요한 문장 부호가 제거됩니다.

더 큰 파일, 더 명확한 플랜

• 무료 업로드 제한 해제：모든 워크스페이스에서 파일당 30GiB 상한을 공유하며, 유료 플랜의 시간 상한은 8시간입니다.

• 게이트웨이 업로드 최적화：대용량 파일은 게이트웨이를 통해 업로드되도록 개선되어, 긴 파일이 마지막 단계에서 실패하는 일이 없습니다.

• 모바일 업그레이드 경험：모바일 플랜 대화 상자가 확장 가능한 카드로 변경되었으며, 사용자에게 적합한 업그레이드 옵션이 미리 확장되어 표시됩니다.

• 크레딧 실시간 표시：크레딧 바에 이번 작업의 비용과 남은 크레딧이 동시에 표시됩니다.

올바르게 표시되는 공유

• 미리보기 카드 최적화：공유 링크를 대화창에 붙여넣을 때 일반 브랜드 카드가 아닌 프로젝트 제목, 공유자 및 미리보기 카드가 표시됩니다.

• 안전한 방문자 권한：방문자는 읽기 전용 에디터로 접속하며, 링크는 문서의 현재 버전만 제공합니다. 수정 기록과 전송 전 삭제된 내용은 보낸 사람에게만 유지됩니다.

프로젝트 시작하기

• 클릭 한 번으로 설정 재사용：프로젝트 목록에서 완료된 프로젝트의 설정을 클릭 한 번으로 가져올 수 있습니다.

• 스마트 언어 정렬：언어 메뉴에 사용자가 해당 단계에서 실제로 선택할 가능성이 높은 언어가 우선적으로 표시됩니다.

• YouTube 전용 다운로드：전용 다운로드 도구가 처리하며, 라이브 스트리밍, 멤버십 전용 또는 일시적인 오류가 발생한 링크의 경우 명확한 원인을 설명합니다.

계정 관련

• Subanana 브랜드 인터페이스：로그인 시 Subanana 브랜드 인터페이스가 적용되며 현지화 연결을 지원합니다.

• 계정 보안：로그인 이메일을 변경할 수 있으며 충돌 발생 시 복구할 수 있습니다.

• 언어 지원：인터페이스에 한국어가 추가되어 영어, 번체 중국어, 일본어와 함께 제공됩니다.

기타 개선 사항

• 텍스트 선택 영역이 테마 색상을 따릅니다.

• 가격 안내 대화 상자가 Safari에서 정상적으로 열립니다.

• AI 채팅이 페이지 새로고침 없이도 자동으로 재연결됩니다.

We spent the last three months rebuilding the subtitle editor around one principle: your edits and your content come first. Everything you type now lands instantly, stays safe through reloads and shaky connections, and exports exactly as you see it.

TL;DR

• Every keystroke, split, merge, and undo now applies instantly on your device and syncs in the background — no more waiting on the network.

• Your text has one source of truth. The old glitches — disappearing characters, text lost while splitting a cue, stale translations — are gone at the root.

• Redesigned preview and Timeline, a Source track switcher for multi-source projects, cross-track Find and replace, dependable Version history, and faster, more faithful exports.

Editing at the speed of typing

The old editor confirmed many edits with the server before they fully settled — on a slow connection, undo after splitting a cue could hang for seconds. The new engine commits every change to your device first and syncs quietly behind the scenes. Typing, splitting, merging, and undo/redo are now immediate, every time, even mid-sync.

Built to never lose a word

Subtitles are your content, and the editor's first job is to keep every word safe. The new engine stores your track as a single authoritative document: edits survive page reloads, offline moments, and flaky networks, and each change is recorded exactly once — no duplicates, no silent drops. The bug classes that used to bite — deleted characters reappearing, text vanishing during a split, translations going stale — were all symptoms of the old architecture, and they're gone with it.

Preview and Timeline, redesigned

Drag a cue on the Timeline and its neighbours now shift out of the way instead of blocking you. Seeking is precise, and playback follow no longer yanks the cue list around while you're editing. The preview stays true to what your export will look like.

Every track, one editor

Projects with multiple sources now open every track in the same editor — switch between them with the Source track selector. Find and replace works across tracks, so one pass can clean up a term everywhere.

Version history you can rely on

Version history snapshots both your content and your track lineup. Restoring a version brings back the project exactly as it was then — including which tracks existed — and nothing is ever truly lost: any version remains recoverable.

Sharper, faster exports

A new rendering engine powers subtitle export. Rendered videos match the editor preview more closely and finish faster.

Why a three-month rebuild?

We could have kept patching glitches one by one. But every fix was fighting the same root cause: the old editor kept several copies of your subtitle text and synced them by hand, and every gap between copies was a chance to lose your work. That's not good enough for the place where your content becomes final. So we rebuilt the editing core on the same collaborative-editing technology that powers modern shared documents — one owner for every character, offline-safe by design. It took three months to do it right, and it's the foundation the next year of editor features will stand on.

This is just the beginning

The new engine unlocks what we hear asked for most — real-time collaboration is now on the horizon. To try the new editor: open any subtitle project. It's already on for everyone.

TL;DR

• New Share button on every project — invite a teammate by name, or turn on Anyone with the link for a public read-only view.

• Pick Can view or Can edit per recipient. Viewers get a clean read-only editor.

• A new Shared with me tab on the dashboard collects every project shared into your account.

• Revoke access (or disable the link) and everyone is kicked in real time.

Invite a teammate, or share a link

Invite by name. Type a teammate's name in the picker — every workspace member is searchable, not just the first page. They show up in the avatar stack on your side, and the project lands in their Shared with me immediately.

Anyone with the link. Flip the scope to Anyone with the link, copy the URL, or open the QR popover for a downloadable PNG. Guests see a read-only preview of the project with a one-click Save to my workspace — on sign-up, the project auto-claims into their account.

View or edit — and a clean read-only mode

Can view — read, scroll, play, and export. A Read-only badge sits in the panel headers; everything that would mutate the project is hidden.

Can edit — cue edits, translations, almost everything an owner can do. Settings and the quality-feedback prompt stay owner-only.

The project Owner is the creator and stays that way in this first cut.

You stay in control, in real time

Remove a recipient and they're bounced off the project with a toast. Disable the public link and every guest who claimed it is kicked in a single sweep. Flip the link role between Can view and Can edit and every active recipient updates in place.

Recipients can also self-remove from Shared with me — kebab menu → Remove from shared.

This is just the beginning

This is our first cut at proper project sharing. We'll iterate fast on how you actually share work and what you tell us — folder-level sharing and per-recipient role editing are up next.

To try it: open any project, click Share in the top right, and either pick a teammate or flip the scope to Anyone with the link.

TL;DR

• New Standard / Enhanced transcription engine toggle in the upload drawer. Enhanced auto-polishes every cue and uses 1.5× credit.

• Polish is rewritten as "diagnose first, apply once" — a ~10–30 s scan surfaces 4–5 categories of suggestions with real before → after examples from your own cues, and only the categories you tick get applied.

• Live progress in the editor — no polling, no manual refresh.

Two transcription engines, your call

Pick Standard or Enhanced in the upload drawer (Enhanced has a NEW badge).

Standard is the fast path — best for rough cuts, sync-checking, and quick turnaround. 1× credit.

Enhanced runs an extra polish pass on top of Standard, auto-cleaning punctuation, line breaks, and common transcription slips. Best for final subtitles you'll publish as-is. Uses 1.5× credit.

The Estimated Transcription Usage shows a clear ×1.5 pill the moment you flip to Enhanced, so the multiplier is explicit before you submit.

Polish, redesigned — "diagnose first, apply once"

The old review-every-cue flow is gone. New shape:

1. Click Polish. The system scans your transcript (~10–30 s).

2. The AI lists 4–5 polish suggestion categories — e.g. Filler-word cleanup, Name spelling, Punctuation normalization — each with a plain-language description, a real before → after example pulled from your own cues, and a count of how many cues it would touch.

3. Tick the categories you trust. Skip anything risky.

4. Click Polish again. The selected fixes apply in one pass, with cues updating live in the editor.

If the scan finds nothing actionable, you'll see an Already polished or Almost perfect screen instead of an empty options list.

This is just the beginning

This is our first cut at smarter, in-product polish. We'll iterate fast on how you use it and what you tell us — consistent, personalized styling control is up next.

To try it: open the upload drawer, drop a file, toggle Enhanced, submit. Or open an existing project and click Polish in the editor toolbar.

Stop fixing the same misspellings, re-explaining your product, and rewriting summaries into your team's shape. Three new sidebar tabs let you teach Subanana once — every project benefits.

Glossary

Pin terms you don't want misspelled (brand names, people, jargon).

• Workspace list applies everywhere; Custom lists attach per project.

• Add one-by-one, paste a batch, or bulk import XLSX/CSV.

• Mark Universal (same in every language) or Language-specific.

Background Pack

Hand the engine the brief upfront. Upload PDF, DOCX, PPTX, XLSX, TXT, MD, paste text, or drop a URL. Auto-classified into 16 reference types and structured into Summary, Key Points, Facts, Key Terms, and Tone.

Template Library

30+ built-in summary templates (Meeting, Speech, Business, Medical, Legal, Creator, and more). One click to use. Build your own with Generate with AI.

Supporting Materials picker

Pick which glossaries, background files, and template apply to a project — works for single files, batch uploads, and live caption (ASR + translation read your context too).

Also shipped

• Smart meeting formatting is now always on.

• Mobile-first redesign of all three tabs.

• Plan limits visible on the pricing card.

• "How it works" popover on every page.

Coming next

Auto-suggested glossary terms from your edits, inline attribution badges, and per-project glossary overrides.

New Features

Live Captions

• Optional language hints in auto-detect mode — When using automatic source-language detection for live captions, you can now provide likely language or script hints to help recognition lock on faster.

• Faster first load for long sessions — Live caption history now opens from the latest part of the session first, so long-running sessions feel much faster to open.

• On-demand loading for older caption history — Older live caption history is now loaded only when you scroll back to it, which makes large sessions easier to browse.

Exports

• Subtitle styling in video exports now matches the editor preview more closely — Video export rendering has been improved so the final output stays true to what you see while editing.

• Distinct file names when multiple tracks share the same language — If you export a project with imported or re-transcribed subtitle tracks in the same language, exported file names and sheets now stay distinct instead of overwriting each other.

Fixes

Sign-In & Access

• Smoother recovery from expired sign-in links — If sign-in returns with an expired session, the app now restarts the sign-in flow automatically instead of leaving you on an error page.

Live Captions & Translation Quality

• Short or mixed-language captions translate more reliably — Auto-translation for difficult short inputs is now more responsive and less likely to stall.

• Live captions handle temporary service interruptions more gracefully — When one upstream provider times out, processing now moves on more predictably instead of repeatedly stalling.

• Fewer awkward short fragments at the end of a caption line — Live captions are less likely to split into one- or two-word tail fragments.

• Captured audio is preserved through reconnects — Live caption sessions now recover more safely from connection interruptions without dropping speech that was already captured.

New Features

Subtitle & Caption Editor

• Find & Replace redesigned — The find-and-replace dialog and header actions in the editor have been rebuilt for a smoother, less cluttered experience.

• Export directly from the project menu — You can now start an export from the project list without opening the full editor first.

• Scene presets for quick project setup — Common scene types are now available as presets when creating a project.

Transcription Accuracy

• Improved transcription accuracy — Several improvements have been made to transcription quality, reducing errors and missed words across a wider range of audio.

• Better subtitle line quality — Subtitles are split into lines more naturally, including improved handling of mixed scripts.

Meetings and Transcript

• Smart Segmentation — You can now choose to group consecutive speech into paragraphs and restore punctuation for better readability.

• Improved meeting editor on mobile — The meeting editor's toolbar collapses overflow actions into a "More" menu on smaller screens.

File Uploads & Import

• Large files upload more reliably — Large files now use a multipart upload with a real-time progress indicator. Files that are still uploading or have failed are blocked from being accidentally selected.

• Clear error for files that exceed the duration limit — Instead of failing silently, the app now shows a clear message when a file is too long.

• Files without audio are handled gracefully — Uploads without an audio track now show a clear explanation instead of failing without feedback.

Project Organisation & Folders

• New projects stay in the folder you are in — Creating a project from inside a folder now keeps it in that folder, instead of moving it to the root level.

• Recently used language variants appear at the top — The language selector surfaces the variants you use most often for quicker selection.

• Search "Chinese" to find Cantonese and Mandarin — Typing "Chinese" in the language selector now shows both Cantonese and Mandarin options.

Live Translation & Live Captions

• Switch your microphone — You can now change your audio input in the live translation admin panel.

• Friendlier errors and auto-reconnect — Live translation now shows clearer error messages and reconnects automatically when the connection drops.

Fixes

Live Translation & Live Captions

• Admin panel history loads reliably — A bug that sometimes showed only partial captions when navigating back to a session has been fixed.

• Captions scroll correctly on iPhone and iPad — Live captions now stay in sync with the latest text on iOS Safari.

• Pending captions survive a pause — Captions that were in-flight when you paused are now properly recovered when you resume.

• More reliable reconnection — The live translation display reconnects more reliably after network interruptions, including on Safari.

• Start and pause work reliably — Fixed cases where the start button could get stuck after pausing, or where pausing during a resume wasn't possible.

Subtitle & Caption Editor

• Timeline drag, trim, and keyboard shortcuts polished — Drag-to-resize, ripple trim, and keyboard shortcuts in the timeline all work more reliably.

• Mobile editor stability — Multiple fixes for playback, menus, and video loading on mobile.

Transcription & Accuracy

• Clearer error messages when transcription fails — Projects that hit a permanent error during transcription now show a clear reason instead of silently retrying.

• Improved transcription accuracy — Additional fixes to transcription output, reducing hallucinations and errors in edge cases.

Meetings

• AI chat in meeting editor works reliably — Fixed a bug where the AI chat would load indefinitely in the meeting editor.

• Calendar recording reliability — Fixed issues with calendar recordings being missed during sync and with errors when connecting a calendar.

File Uploads & Import

• iPhone Voice Memos and M4A files upload correctly — M4A files from iPhone, including Voice Memos, were being rejected. This is now fixed.

• Audio-only files work end-to-end — Projects created from audio-only files (MP3, M4A, podcasts) now work correctly throughout the entire workflow.

• Large video uploads no longer block the browser — Uploading large video files no longer causes the browser to become sluggish or unresponsive.

Exports

• ZIP files open on Windows — Exported ZIP archives now unzip correctly on Windows without errors.

Onboarding & Sign-In

• Faster landing after sign-in — The check after sign-in has been streamlined so you reach the dashboard without an extra redirect.

We have spent months completely rebuilding the entire app architecture, addressing previous pain points in your workflow and introducing a brand new user interface.

--------------------------------------------------------------------------

Brand New Workspace

We've redesigned the most commonly used interface to make operations more intuitive:

• Subtitle Editor: Bilingual subtitles, basic video editing, and video export preview are finally integrated into a single workspace, making editing smoother.

• Meeting Editor: Features a dual-column layout with AI highlights on the left and transcript on the right—no more frequent page switching. Added an AI note template library:
• Extensive Preset Templates: No need to rack your brain writing prompts—choose from a variety of built-in summary formats.

• Custom Templates: Just provide a document sample and AI will mimic the tone and format to generate content.
• Smarter Ask AI: Improved Q&A experience—want to know "Who mentioned the budget in the meeting?" Just ask AI for more accurate answers.

• AI Spell Check: Enhanced automatic typo detection and correction—easily review and apply fixes.

• Operate with Confidence: Supports Undo/Redo, so you can easily revert mistakes.

Faster and More Stable

• We've updated the underlying architecture to reduce loading times and errors.

Clearer Calendar Meeting Management

• The new page better integrates your calendar. You can manage all meeting recordings and transcription settings here.

--------------------------------------------------------------------------

🔮 Coming Soon: Faster Iteration and New Features

This update will greatly accelerate our pace of developing new features. In the coming quarters, we will be rolling out:

• More Flexible Plans: Updated pricing and plan options to better match your actual usage.
• AI Smart Line Breaks: No more manual adjustments! Customize characters per line—AI will automatically analyze tone and grammar to re-break lines.
• AI Agent Assistant: More than just "recording"—AI will proactively help you complete tasks (such as sending follow-up emails, syncing Notion or Slack project updates).
• Seamless Sharing & Collaboration: Easily share folders, meeting notes, and video subtitles with your team or external clients for collaborative work.

--------------------------------------------------------------------------

We can't wait for you to experience the all-new Subanana.

If you have any questions about this update, feel free to reply to this email and contact us.

• Support linking Microsoft Outlook Calendar to set up automatic meeting recording
• Improve and update some page designs
• Fix some operations in the editing interface and improve smoothness
• Improve AI summary generation content

[Real-time Translation]

Added AI Real-time Translation feature for generating real-time subtitles and translations for meetings and events

After setting up real-time translation, you can start recording, and the system will instantly display transcription and translation results side by side

* Exclusive feature for Ultra and Business users, but all users can try this feature

Feature highlights:

• Side-by-side bilingual subtitles: Allow all participants from different backgrounds to engage more in discussions
• Customizable vocabulary style: Ensure translations accurately reflect industry terminology, company/product names
• View subtitles on mobile: Use your personal device to scan the QR Code and view real-time subtitles and translations
• Content security and confidentiality: Not used for any AI learning, and content can be deleted at any time
• Export text records: View records and download subtitle files anytime, making post-production easier

The AI real-time transcription tool is suitable for various industries and scenarios:

• 🏢 Seminars, Webinars, and large-scale events
• 🎓 Real-time translation for classes
• 🏥 Medical diagnosis and consultations
• 🎤 Media interviews, Podcasts
• ✈️ Travel tours

[Upload Video/Audio] Users can now select [Record Page Content] and upload transcribed subtitles or meeting notes

*Available only on MacOS and Windows recording

*You must record a specific window to capture audio content

[Upload Media] Users can now select [Send email notification after project completion] at the bottom right of the upload page. After all transcriptions are completed, users will receive a completion notification at their account email.

[Meeting Notes] Users can now select different AI models to assist in generating content when using AI generation features (AI Summary, Chapters, Ask AI)

[Meeting Notes] Improved the layout of text content in the editing page to make it clearer and easier to read

[Meeting Notes] Removed the text preview below the video in the editing page

[Meeting Notes] When connecting to Google Calendar, if a scheduled recording is changed in the calendar, the scheduled recording time will also update automatically

[Meeting Notes] Fixed and improved some text descriptions and navigation in Google Calendar

[Transcription Feature] If the uploaded content fails to transcribe, users can now click the [Retry] button on the right to attempt transcription again.

[Transcription Feature] Upload hours will now be deducted after a successful upload, instead of after a successful transcription, to prevent users from over-uploading and causing eventual transcription failures.

[Account] Users can now delete their Subanana account by themselves if needed.

You can delete your Subanana account by following these steps:

1. On the project list page, click [User Avatar] at the top right -> [Account]
2. Under Personal Information, in the Delete Account section, click [Click Here]
3. Follow the instructions to confirm the information, then click [Delete Account]

Please note! Once your account is deleted:

• Your paid subscriptions and usage will be cancelled and will not be refunded.
• All audio and video files you have uploaded will be deleted and cannot be accessed again.
• All transcription content you have generated will be deleted and cannot be accessed again.
• Your account and personal information will be deleted.
• This action will take effect immediately and cannot be undone.
• You will not be able to use Subanana again unless you re-register.

All users can now share a unique link to refer friends to subscribe to the Subanana plan.

Friends who subscribe through this link can receive a subscription discount, and users will also receive corresponding rewards for successfully referring friends.

Click the gift button on the right side of the page to get your unique link:

Meeting notes will now support exporting with reference times when generating chapters and summaries, allowing users to find relevant content more quickly.

[Meeting Notes] Added the feature to link Google Calendar for automatic scheduling and recording of Google Meet meetings

• You can now link Google Calendar and sync it to Subanana to set up automatic recording schedules for future Google Meet meetings

--------------------------------------------------------------------------

How to use

1. On the right side of the meeting notes page, click [Calendar], then click [Connect Google Calendar]

2. After logging in, check all items on the right on the Google confirmation page, then click [Continue]

3. Select the calendar you want to integrate and link, then click [Confirm]

4. In the calendar, you can check meetings as recording schedules as needed

5. You can see the scheduled meeting recordings under the [Meetings] section

--------------------------------------------------------------------------

Others:

[Meeting Notes] Improved translation output for meeting notes

[Meeting Notes] Meeting notes summary AI now provides more accurate and detailed generated content when outputting timeline-related information