更新
每个 Lumenwerk 工具当前的版本，以及每次发布的变化。
Lumenwerk 正在开发中。低于 1.0 的版本为预览版；1.0 代表工具正式公开发布。
当前版本
发布说明
0.3.0
标题与关键词AI最新
YouTube chapters, in step with zooms
Your chapter titles become a list to paste into the YouTube description, and titles and zooms no longer get in each other's way.
- 新增Copy a ready YouTube chapter list (0:00 Intro and every chapter) with one click.
- 改进Keywords prefer the moments the picture is zoomed in, and a callout during a zoom keeps clear of the bigger face.
0.4.0
Auto ZoomAI最新
On its own, and around your titles
Run Auto Zoom on a single file without opening the editor. Zooms now stay out of chapter cards, lower thirds and list titles.
- 新增Standalone: drop a file, preview the zooms on the original, export or open the result in the editor. Subtitles and Studio Sound work the same way.
- 改进No punch-in while a title is on screen, and zooms prefer the moments a keyword pops, so the two land as one beat. Lumi plans it the same way.
0.8.0
Lumi ArtistAI最新
Pick your AI model
Ten of the strongest AI models to choose from, right where you start the AI. Auto keeps picking the right one for each step.
- 新增A model menu: Claude, GPT, Gemini or Grok, grouped by smartest, strong and balanced, fast and cheap.
- 新增The price follows the model you pick and shows before every run, as always.
查看更多 · 还有 1 项变化收起
改进
- Auto stays the default: each step runs on the model that does it best.
0.6.0
Autocut最新
Pick your AI model
Ten of the strongest AI models to choose from, right where you start the AI. Auto keeps picking the right one for each step.
- 新增A model menu: Claude, GPT, Gemini or Grok, grouped by smartest, strong and balanced, fast and cheap.
- 新增The price follows the model you pick and shows before every run, as always.
查看更多 · 还有 1 项变化收起
改进
- Auto stays the default: each step runs on the model that does it best.
0.3.0
Auto ZoomAI
Pick your AI model
Ten of the strongest AI models to choose from, right where you start the AI. Auto keeps picking the right one for each step.
- 新增A model menu: Claude, GPT, Gemini or Grok, grouped by smartest, strong and balanced, fast and cheap.
- 新增The price follows the model you pick and shows before every run, as always.
查看更多 · 还有 1 项变化收起
改进
- Auto stays the default: each step runs on the model that does it best.
0.4.0
AutoScaleAI最新
Pick your AI model
Ten of the strongest AI models to choose from, right where you start the AI. Auto keeps picking the right one for each step.
- 新增A model menu: Claude, GPT, Gemini or Grok, grouped by smartest, strong and balanced, fast and cheap.
- 新增The price follows the model you pick and shows before every run, as always.
查看更多 · 还有 1 项变化收起
改进
- Auto stays the default: each step runs on the model that does it best.
0.2.0
标题与关键词AI
Pick your AI model
Ten of the strongest AI models to choose from, right where you start the AI. Auto keeps picking the right one for each step.
- 新增A model menu: Claude, GPT, Gemini or Grok, grouped by smartest, strong and balanced, fast and cheap.
- 新增The price follows the model you pick and shows before every run, as always.
查看更多 · 还有 1 项变化收起
改进
- Auto stays the default: each step runs on the model that does it best.
0.2.0
字幕AI最新
Pick your AI model
Ten of the strongest AI models to choose from, right where you start the AI. Auto keeps picking the right one for each step.
- 新增A model menu: Claude, GPT, Gemini or Grok, grouped by smartest, strong and balanced, fast and cheap.
- 新增The price follows the model you pick and shows before every run, as always.
查看更多 · 还有 1 项变化收起
改进
- Auto stays the default: each step runs on the model that does it best.
0.7.0
Lumi ArtistAI
Clearer conversations
Lumi Artist sums up what it understood and asks one clear question at a time, knows brand and genre looks, and rebuilds a scene whenever you ask for a change at a given second.
- 改进Replies come as a short summary plus a separate question, with the important part highlighted.
- 改进Ask for a change at a specific second and that scene is rebuilt, never handed back unchanged.
查看更多 · 还有 2 项变化收起
新增
- Name a brand or a genre and the graphic follows its look.
- Before the build starts, one follow-up question with examples fills the gaps.
0.5.0
Autocut
Tighter pauses, easier preview
Kept pauses stay short, silence hidden behind stretched words goes too, and the preview loops and scrubs.
- 改进Pauses Autocut keeps on purpose are never longer than 1.2 seconds.
- 修复Silence after a stretched word no longer slips through the cut.
查看更多 · 还有 2 项变化收起
新增
- After the cut, apply another tool straight away.
改进
- The preview starts again from the top when it ends, and you can scrub by dragging.
0.2.0
Auto ZoomAI
AI picks every zoom
Auto Zoom now always lets the AI read your video and choose the moments. It works in any language you speak.
- 改进The AI reads the transcript and picks what deserves a zoom; the built-in rules only fill in around it.
- 改进Zooms for videos in Spanish, Japanese and every other language, not only English and German.
查看更多 · 还有 1 项变化收起
改进
- The switch for the AI is gone. The price shows before every run, as with every AI tool.
0.4.0
Autocut
Autocut on its own
Use Autocut without the editor: drop in a video, pick what to cut, run it and export. Or open it in the timeline as before.
- 新增Tools > Autocut asks how you want to use it: on its own or in the timeline.
- 新增On its own: upload, settings, run, then straight to the export (MP4, audio, or a project file for Resolve, Premiere and Final Cut).
- 新增Watch the result before you export: the preview plays the cut version (or the original to compare), and you can untick any cut right there.
查看更多 · 还有 1 项变化收起
新增
- Want to check the cuts? Open the result in the editor with one click.
0.6.0
Lumi ArtistAI
A motion graphic as a block
Drop a motion graphic block onto the timeline like a text. Lumi asks what it should show, then builds it right there, as long as the block.
- 新增Motion graphic blocks in the editor's Text panel: click or drag one onto the timeline, tell Lumi what it should show, press Build. The finished graphic takes the block's place.
- 新增Choose where it goes: over your picture, full frame with you small in a corner, or full frame on its own. The conversation is free; you see the price before it builds.
0.3.0
AutoScaleAI
Scale by hand
A Scale block puts you small in a corner wherever you drop it, without any insert or AI. Free.
- 新增Scale blocks in the editor's Text panel: click or drag one onto the timeline and the speaker below shrinks into a corner for as long as it runs. Corner, size and look are in the inspector.
0.2.0
AutoScaleAI
A layout you can edit
Where the speaker sits is now its own lane in the editor. Back-to-back inserts keep the speaker small, a new corner is reached with a glide, and five picture-in-picture looks come with your style.
- 改进Back-to-back inserts no longer make the speaker grow and shrink in between. When the next insert wants another corner or size, the speaker glides over; next to a full-frame insert it moves while it is covered.
- 新增A layout lane above the speaker's track shows when they sit small and where. Change corner, size, look or picture-in-picture versus full frame in the inspector. Layouts move, trim and go with their insert.
- 新增Five looks for the small speaker: Classic, Card, Square, Portrait and Bubble, with border and shadow. Square, portrait and bubble frame the face. A style can set the look for AutoScale and Lumi.
查看更多 · 还有 2 项变化收起
新增
- Right-click a clip: “Shrink speaker” puts the speaker in a corner while it plays, “Picture-in-picture here” does it on the speaker's own clip.
修复
- Exports to DaVinci Resolve, Premiere and Final Cut from the editor keep the speaker's picture-in-picture moves. Picture-in-picture shapes (rounded corners, circle, border) only show in the browser render.
0.5.0
Lumi ArtistAI
Templates: polished and free
Titles, numbers, charts, steps, quotes, name tags and more now come from hand-built templates. They build instantly and cost no credits; only unique scenes are built from scratch.
- 新增Eleven templates for the graphics every video needs: title, big number, bar chart, steps, logo, quote, before/after, lower third, single word, timeline and word slam. Lumi picks them while planning.
- 改进Template scenes are free. The price before building only counts the scenes Lumi builds from scratch.
- 改进Templates fit 16:9, 9:16 and 1:1, shrink long text so it always fits, and sit on a readable plate over your footage.
0.4.0
Lumi ArtistAI
Ideas, taste and a final polish
Lumi turns your request into a real concept, builds with eleven fonts and a library of proven techniques, and looks at every finished scene to fix what doesn't hold up.
- 新增The idea first: before planning, Lumi writes a concept with one big idea, a motif that runs through the video and a signature moment. You see it in the scene plan.
- 新增Fine-tuning: after the build Lumi looks at every scene as a picture, like an art director, and fixes empty frames, clipped text or a weak focal point. You can switch it off before building.
- 改进Eleven fonts instead of two, from tight tech grotesks to wide display faces, an elegant serif and a monospace, plus kinetic type with masked word and letter reveals, drawn lines, counters, wipes and camera moves.
查看更多 · 还有 1 项变化收起
修复
- A scene that crashed while rendering could end up broken in the video; the render now stops and refunds your credits instead.
0.3.0
Lumi ArtistAI
One film instead of separate scenes
Before building, Lumi designs the whole video: shared type and components, a background that runs through every cut, and objects that carry from one scene into the next.
- 新增Match cuts: a logo, a card or a word stays exactly in place across the cut and moves on in the next scene. The preview checks every cut and offers a fix if something jumps.
- 改进Scenes share one look: the same type sizes, cards and colors, plus a background layer that keeps moving through the whole video.
- 改进Bolder pictures: big type, full-frame color, particle fields and drawn curves, and Lumi picks a length that fits instead of a short default.
查看更多 · 还有 2 项变化收起
改进
- A scene now costs 32 credits, and a video of several scenes 39 credits more for its design. The price shows before you build.
修复
- An animation in one scene could move an element with the same name in another scene.
0.2.0
Lumi ArtistAI
Shorter chat, sound in the preview
Lumi says less and marks what matters, you hear music and effects while you watch the preview, and the voice-over now always ends up in the render.
- 改进Lumi answers in one or two short sentences and marks the words that matter in orange. The scene plan and the details fold away until you open them.
- 新增The preview plays music and sound effects in sync, and a finished render plays right in the chat with every sound, voice-over included.
- 新增Attach your logo or screenshots with your first message. Lumi uses them and doesn't ask again.
查看更多 · 还有 2 项变化收起
修复
- Switching on the voice-over after planning left the render silent; Lumi now writes the lines right away.
- After describing your own style, the style suggestions no longer come back, and your answers to Lumi's questions about files show up in the chat.
0.1.0
Lumi ArtistAI
Describe it, Lumi animates it
Lumi Artist turns a description into a motion graphic: a lower third, a logo sting or a whole short promo, planned with you in a chat and rendered with voice-over, music and sound.
- 新增Say what you want in your own words. Lumi asks only for what's missing, plans the scenes with you or lets you plan them, and shows the full plan with its price before anything is built.
- 新增Play it right away, then ask for changes in the chat. Only the scenes you change are made again, and they keep everything you didn't ask to change.
- 新增Render as MP4, or as a transparent overlay (WebM or ProRes 4444) to lay over your own footage, and put it straight into an editor project.
查看更多 · 还有 3 项变化收起
新增
- Voice-over in several voices, music in five moods with a drop on the right scene, and sound effects on the cuts. Mixed to −14 LUFS like the rest of your video.
- Four starting styles, or describe your own look, colours and fonts. Your logo and screenshots go right into the graphic.
- Sound changes need no rebuild: change the voice-over or the music and render again. If part of the animation breaks in the preview, Lumi can fix just that part.
0.1.0
标题与关键词AI
Titles and keywords, right on the word
The AI reads your transcript and decides where a chapter title, a name or a keyword belongs, writes it and puts it on the timeline.
- 新增Chapter titles when a new topic starts, lower thirds when someone is introduced, numbered titles for lists and the occasional key statement.
- 新增Numbers, names and key terms pop up next to you the moment you say them, on the side away from your face.
- 新增Check everything before it goes in: untick what you don't want and fix the wording right in the list.
查看更多 · 还有 4 项变化收起
新增
- Display times follow reading speed, chapters stay at least a minute apart, never two graphics at once, and nothing lands on words Autocut removed.
- Three looks (clean, bold, editorial) and three keyword densities. Every title stays an editable text clip.
- Without the AI, free rules still find numbers, lists, announced topics and a speaker introducing themselves.
改进
- Text clips can now sit left or right and have a smaller second line. SRT exports leave titles and keywords out.
0.1.0
AutoScaleAI
Your screen, right when you talk about it
AutoScale watches your screen recordings and other inserts, puts each part where you talk about it and shrinks you into a corner while it plays.
- 新增Knows what happens in a screen recording: new pages, typing, clicking, scrolling and idle stretches. Idle parts are left out, small text is zoomed into.
- 新增Recorded your screen and camera at the same time? AutoScale lines them up by the sound and shows the screen whenever something happens on it.
- 新增Recorded separately? Your announcements (“let me show you”) place it, or the AI looks at a few stills and matches each page to what you say.
查看更多 · 还有 2 项变化收起
新增
- The speaker goes into the corner where nothing happens on screen, with a close-up of the face. Footage and images get the full frame. Smooth or hard transitions.
- You check every insert before it goes in. Exports to Premiere, DaVinci Resolve and Final Cut as moves and keyframes you can still change.
0.3.0
Autocut
Cuts like an editor
After transcription, an AI reads the transcript like an editor and decides what goes. Autocut still places every cut on a clean point in the audio.
- 新增Retakes are found even when the second take is worded differently. The last complete take stays.
- 新增Abandoned sentences and remarks like "wait, again" are cut. Words like "basically" or "halt" only go where they add nothing.
- 改进Every AI cut shows why it was made. Cuts the AI did not confirm stay in the list, switched off.
查看更多 · 还有 1 项变化收起
改进
- The AI review is saved with the project, so reopening it costs nothing. Only the transcript text is sent, never audio or video.
0.3.0
Studio Sound最新
One audio tool, three modes
Noise Removal and Dialog Leveler are now modes of Studio Sound: "Full", "Denoise" or "Level".
- 新增"Denoise" takes out noise, hum and pops and leaves level and tone as they were, for audio you mix yourself.
- 新增"Level" brings quiet sentences up and loud ones down by up to 15 dB and hits your loudness target, without touching noise or tone. Pauses are never pushed up.
查看更多 · 还有 1 项变化收起
改进
- Projects and links made with Noise Removal or Dialog Leveler open in Studio Sound in the matching mode, with their settings.
0.2.0
Studio Sound
Quiet voices stay clear
A new noise reduction keeps far more of a quiet or distant voice, a quiet speaker now ends up at the same level as the rest, and the room echo reduction measures your room instead of guessing.
- 改进A voice that is barely louder than the noise keeps its level and clarity: about 3 dB lost instead of 7 in our tests, with the same noise reduction and no chirping artifacts in the pauses.
- 改进A speaker 15 dB quieter than the rest now ends up within about 1.5 dB of it (before: about 4 dB), and the background in their pauses is no longer pushed up with them.
- 新增Pop removal: the thump of p and b sounds on a close mic is turned down for its few milliseconds, the rest of the voice is untouched.
查看更多 · 还有 2 项变化收起
改进
- Room echo reduction measures how long your room rings and adapts to it. A dry room is left alone, a bare room gets more. "What we found" shows the result.
- A fan or air conditioning that gets louder while you talk is followed, so it does not come back in the pauses.
0.1.0
Studio Sound
Studio Sound is here
One click turns a home recording into a clean, even voice at the loudness YouTube expects. Runs in your browser, nothing is uploaded.
- 新增Removes background noise, mains hum and room echo, balances the tone of the voice and tames sharp s sounds, in one pass.
- 新增Hits your loudness target exactly (YouTube -14, podcast -16 or -19, broadcast -23 LUFS) with a true-peak limit, measured on the finished file.
- 新增Listen before you export: A/B between original and processed at the same loudness, plus a loudness curve of the whole video.
查看更多 · 还有 2 项变化收起
新增
- Light, Balanced or Strong with one click, or every step on its own. "Keep some original" brings back a natural sound if it feels too processed.
- Exports a WAV with exactly the length of the video, ready to lay under the clip in Premiere Pro, DaVinci Resolve or Final Cut Pro. Free, no credits.
0.1.0
Auto ZoomAI
Zooms that follow what you say
Auto Zoom reads the transcript and punches in where editors would: on numbers, turns, takeaways, questions, lists and punchlines.
- 新增Finds zoom moments in what is said, in English and German: numbers, “but” and “the truth is”, “the key is”, questions to the audience, “first, second, third” and the line before a laugh. Every zoom says why it is there.
- 新增Each moment gets the right move: a hard punch-in on the word, a smooth zoom, a slow push over a key thought, or one step closer per list item.
- 新增Live preview in the player and export to Premiere Pro, DaVinci Resolve and Final Cut Pro, with hard zooms as edit points and smooth ones as keyframes.
查看更多 · 还有 3 项变化收起
新增
- Pacing calm, standard or energetic, with spacing between zooms so they never feel random. Moments that were left out are offered as alternatives.
- Zooms go toward the speaker's face, found locally in your browser. The framing stays as you shot it, and zooms never go further than your footage allows.
- Builds on your Autocut edit: zooms on removed words disappear, and jump cuts alternate between wide and close like a second camera.
0.1.0
字幕AI
Subtitles are here
Captions from your transcript in every common style, from classic subtitles to one word at a time.
- 新增Show classic two-line subtitles, short phrases or one word at a time. Line length and line breaks follow broadcast rules.
- 新增Important words are marked automatically, or by AI on request, in colour, on a box or bigger. Click any word to change the pick.
- 新增Export a transparent animated overlay (WebM), styled titles for DaVinci Resolve and Final Cut Pro (FCPXML), or SRT, VTT and ASS files.
查看更多 · 还有 3 项变化收起
新增
- Reveal the whole caption at once, as karaoke with the spoken word highlighted, or word by word as it is spoken.
- Optional animation: fade, pop or slide.
- Eight starting styles, with font, size, colour, outline, shadow, box and position all adjustable.
0.2.0
Autocut
Natural pacing
Edits breathe like a person talks, and cut points get cleaner. Autocut now listens to where speech really starts and stops instead of trusting the transcript's timestamps alone.
- 改进Sentence ends keep a slightly longer pause (0.4 s) than joins mid-sentence (0.25 s), so the edit no longer sounds breathless.
- 改进Pause edges follow the real start and end of speech in the audio, also when the transcript's timestamps are off.
- 新增Coughs, throat clearing, sniffs and lip smacks are cut as their own kind. Laughter always stays.
查看更多 · 还有 7 项变化收起
新增
- Finds pauses the transcript hides, where a word's timestamp runs on into the silence after it.
- Option to keep the breath before a new sentence. By default the breath goes out completely instead of being cut in half.
- Fillers glued into running speech, where a cut would click, stay suggested but start switched off.
- Suggests abandoned sentence starts that break off mid-word, switched off by default.
改进
- More of the kept pause sits after a word than before the next one, so word endings and reverb tails aren't clipped.
修复
- A cough inside a pause no longer splits it into two pauses that are each too short to cut.
- "dass das" and similar word pairs are no longer flagged as stutters.
0.1.0
Autocut
Autocut is here
Autocut finds the parts of a recording nobody needs to hear and suggests them as cuts you review one by one.
- 新增Finds long pauses, filler words like "um" and "uh", stutters and retakes. For retakes, the last take stays.
- 新增Every cut can be switched on or off in a review list, with transcript and timeline side by side. Uncertain cuts start switched off.
- 新增Export as a project for Premiere Pro, DaVinci Resolve or Final Cut Pro (XML / FCPXML) plus SRT subtitles.
查看更多 · 还有 4 项变化收起
新增
- Pauses are shortened to a natural pause instead of a hard jump.
- Filler words in 90+ languages, plus optional soft fillers like "basically" or "halt".
- Cut edges snap to the quietest nearby frame, so no syllable gets clipped.
- The video stays on your computer. Only the extracted audio track is sent for transcription.