Tutorials

The last step is Share. First, connect your profiles under the Accounts tab, Instagram, TikTok, YouTube, LinkedIn, or Facebook. Then pick which accounts to post to, and either Post now or Schedule post for a date and time. Everything you schedule shows up on the Calendar, across all your projects, where you can see and reschedule at a glance. And if you'd rather post by hand, you can always download the finished clip instead.

The Caption step writes your post copy. Pick a platform, Instagram, TikTok, YouTube, LinkedIn, or Facebook, and ScriptCut drafts a caption in that platform's tone. Edit it however you like, or hit Rewrite for a fresh take, and Copy to grab it. It's built from the words in your clip, so it's always on message. Captions are a ProAI feature and use one AI credit each.

The Cover step is your thumbnail. Four ways to make one: Pick a frame straight from the clip, Frame + AI to restyle a frame, Generate a fresh image from a prompt, or Upload your own. Choose Textless for a clean image, or With text to have AI render a headline right onto it, and set the tone from conservative to creative. Every image you make is saved to a gallery for this clip, so you can compare and reuse. Covers are a ProAI feature and use a few AI credits each.

Social video lives or dies on captions, so the Subtitles step burns them right into the clip. Pick a style from the preset gallery, clean, karaoke, word-by-word, and more, each with a live preview. Change the color to match your brand, and drag the subtitles anywhere in the frame. What you see in the preview is exactly what renders into the final video.

When a clip needs to go vertical, the Video step handles it. Pick your aspect ratio, nine-by-sixteen for Reels and Shorts, square, or four-by-five. Turn on smart reframe and ScriptCut keeps the speaker centered as they move. Below is a real timeline: scene cuts you can adjust, trim handles to extend or tighten the ends, and Command-scroll to zoom in for fine work. When the framing looks right, click Next.

A quick guided import gets you rolling: upload your video, then let ScriptCut transcribe it or bring your own transcript. Now the clever part. ScriptCut reads the whole recording, works out how many genuinely strong clips it can actually support, and lays the candidates out for you. Go down the list, Keep the winners, Discard the rest. Happy? Hit Accept All, and everything you kept lands on your Social dashboard as its own little story, ready to polish. You'll see the AI credit cost before it runs.

There are two ways in. Pick Social Clips when you start a new project, or hit Social on the Review page of any video project. Either way, the publishing suite opens. Across the top: Clips, your grid of short clips, filed into folders and searchable just like your projects, Calendar, where scheduled posts live, and Accounts, where you connect your profiles. Open any clip and ScriptCut walks you through five steps, ticking a green check as you finish each one: Video, Subtitles, Cover, Caption, and Share. Let's run through them.

The Premiere roundtrip works the same way. Export Timeline (XML) from ScriptCut, then in Premiere Pro go to File, Import, and choose the XML. Your selection arrives as a sequence, in order and trimmed, with your groups as markers. ScriptCut preserves Premiere's own timeline metadata through the roundtrip, so the sequence drops in clean and you skip the entire rough assembly.

When your cut is ready, the Export menu sends it out. Remember the filter, All, Must have, or Nice to have, decides what's included. Choose Timeline (XML) and ScriptCut rebuilds your entire selection as a real timeline: in order, trimmed to the frame, with your groups as markers. Your rough cut is already assembled, matched to your original footage, ready for you to finish.

If you want ScriptCut to cut your timeline for you and you did not not import your raw clip for transcription, you will need to add a XML file. Don't worry, this step is as easy as a click. In Davinci Resolve, choose the timeline you transcribed earlier. This part is important as the timecode has to match your transcript so they both speak the same language. Whether you built a single layer, multi layer, or multicam timeline, doesn't matter. ScriptCut will work roundtrip either. Select your timeline, right click, chose Timelines, Export, and find XML. Rename your file if needed and click on the dropdown menu. Choose FCP 7 XML V5 Files. Import the file with the .xml extension and you're done. Once you finished your work in ScriptCut, simply export the XML file and return to Davinci Resolve. In your Media Pool, right click, Timelines, Import and choose XML. Alternatively, click Command, Shift i. If you're still in the same project, you can uncheck "Automatically import source clips into media pool". Click OK and you have a ready to go assembled timeline that includes all your notes as markers as well as Marker ranges indicating a group with your group name. This step is optional but worth it. If you export a timeline XML from the same sequence as your transcript, ScriptCut can build the rough assembly for you in your video editor. Export an FCP7 XML of your timeline in your preferred Editing Software. You can export single layer, multi layer, nested or multicam sequences. Just make sure the transcript and the XML come from the same timeline. ScriptCut checks for you, and warns if they don't match. If you imported a Video or audio file and had it transcribed by ScriptCut, you can skip this step and you'll still be able to export an xml that builds the timeline for you, as long as you're using the same raw footage that you transcribed.

To get a word-level transcript out of DaVinci Resolve, open the Audio Transcription panel, click the three-dot menu, and choose Export Subtitles. Now the settings that matter: set Maximum to 1 character per line, set Lines to Single, and check Include Speaker Name. That one-character-per-line setting is the whole trick. It gives every word its own timecode, which is what lets ScriptCut cut to the exact frame. Save the file, then drop it into ScriptCut's transcript step. As of 2026, the transcription engine in Davinci Resolve isn't perfect. Should you be unhappy with the results or experience timing issues, try to Transcribe within ScriptCut for more accurate results.

Creating word level timed json transcripts in Premiere Pro is incredibly easy. To automize the process, navigate to Preferences, Media Analysis and Transcription, enable Automatically transcribe clips, and under speaker labeling select, "Yes, separate speakers." Don't want to enable the automation? No problem. Once a video or audio file was imported, navigate to Text. If you can't find it, click on Window, and check Text. Click Transcribe. Choose your language and under Speaker labeling select "Yes, separate speakers". Once the transcript has processed, click the 3 dot menu, export, export transcript.

Choose Audio and ScriptCut splices just your selected moments into a single file, using your word timecodes to keep only the spoken parts. You decide the feel: Keep natural pauses leaves the breaths between sentences, exactly as spoken, for one continuous clip. Remove gaps cuts the silences for a tight, back-to-back edit. It's the fastest way to hand off a radio cut or a scratch track and perfect for Podcast edits straight out of ScriptCut

Choose Subtitles and you get a proper subtitle export, not an afterthought. Pick your format, SRT or WebVTT, then a standard: Netflix, YouTube, Broadcast, US Caption, a Social vertical style, or your own custom rules. Turn on Split long captions and ScriptCut breaks each sentence into short, subtitle-sized cues that fit the standard you picked, just like Resolve does. It even tells you how many cues already fit before you export.

Not everything needs to open in an editor. Formatted PDF gives you a polished, branded script document, perfect for clients and producers. When working with marketers, they can use the approved transcript pdf to get started on their marketing plan while you're finishing the edit. Editor cut sheet is a plain-text rundown of every selected line with its timecodes, made for whoever's finishing the edit and us helpful to be used as a reference for 100% certainty. Both carry your editor notes and your chapter headings, so the intent travels with the cut.

Here's what your client sees. First a quick step asking who's reviewing, so every note is tracked to a name. Their user icon will show up in the right corner so you can see if your client is currently active. If you uploaded your personal branding, the page will look like your own. Your client can make selects, arrange it, play it back, and leave notes on any line. When they're happy, they click Approve. Approval locks the project so nothing changes by accident, and it tells you straight away, in your Activity feed and by email. If you need to change something after, you can unlock it. Under Settings - Client view, you can give or deny access to certain tools by switching the toggle. By default, client have access to your AI tools that will use your AI Credits. See that chat bubble in the corner? It's not only your and your client's personal AI assist but also a chat between you, your team, and your client, so every question about the project lives in one place instead of scattered across a dozen emails. Download the whole conversation any time.

Below the script is an Editor notes field. Anything you write here travels with the project into the editor cut sheet and the PDF. It's your single place to leave overall direction for whoever finishes the edit: the tone, the must-keeps, the do-not-touch. One note, and it's on every hand-off.

Review presents your selection as a clean, readable script, exactly what your client will see. The total runtime sits at the top, and each group shows its own. To listen or watch the video, just hit play. You can also click on specific sentences to skip around. The filter buttons let you read All selects, Must have only, or Maybe's only. That filter matters: it's the same one that decides what each export contains. So if you export on Must have only, that's what lands in your timeline.

When you attach audio or video, you can play your cut without leaving ScriptCut. Every line has a play button, and you can select multiple rows to play them back to back. If you added video, a floating player opens that you can drag, resize, minimize, or send full screen. Got some issues with your internet connection? Hit Turn video off for lighter, audio-only playback. The player runs along the bottom with a scrubber, volume, and playback speed. And it's always trim-aware: if you trimmed a line to six words, you hear those six words, never the parts you cut. What you hear is exactly what you'll export.

Already made your picks and just want a little headstart with the order? AI Arrange takes the lines you've selected and sequences them into a coherent story, and it can group them into titled acts for you. It understands the full context of your transcript and your actual selections to make the best judgement calls based on hundreds of hours of training. To take more control, visit the advanced settings. As always, the more input you give the AI, the closer the results will match your goal.

Groups are incredibly useful to stay organized and to give your editor a map. Drag a box across a run of lines, right-click, and choose Group selected or click the group icon on top to gather them into a named section: an act, a beat, a chapter, mini social videos. Name it whatever it's about. Those groups carry into your editor as named timeline markers, and into your PDF and your client's review as chapter headings. A timeline that arrives already broken into cold open, turning point, and payoff is far faster to finish than a flat string of clips. Need to start over? Reset arrange clears the groups and order while keeping your picks.

Arrange shows only the lines you selected, and lets you put them in the order that tells the best story. Grab any line by its handle, the icon on the left, and slide it up or down until it flows right. This is where a pile of good moments becomes a story with a beginning, a middle, and an end.

AI Story goes all the way. In a single pass it designs the finished edit: which lines to keep, the order they play in, the sentence-level trims, even an optional cold-open hook. When it's done, it drops you on the Arrange page with the whole story laid out, ready to refine. Like every AI tool, it works from the same brief. And it quietly trims filler and false starts as it builds, so what it hands back already reads clean.

On the ProAI Plan, AI Select reads your whole transcript and marks the strongest moments for you. First you set a short brief: whether you want Short Form, one tight story, or Long Form, a multi-topic selection with story archs. Choose your goal between convert, educate, or entertain, a target runtime, and an optional prompt for anything specific. The AI engine has hundreds of hours of training but as always, the more specific you can be here, the better results you're gonna get. It selects in place, right on the Select page, so every pick is still yours to adjust. Think of it as a strong first pass that saves you an hour, not a replacement for your taste.

Every line has a note icon. Leave a note for yourself or your client, a direction, a question, a reminder, and it travels with that moment all the way into your editor, as a timeline marker. It's how you hand off intent, not just clips. Use Find, or Command-F, to jump to any word in the transcript, and the full-screen button when you want a clean reading surface for a long first pass. Here are few more useful features you should know about. If you know what you're looking for, use Find, or Command-F, to jump to any word in the transcript. Sometimes the UI can feel distracting, so we built a full screen view. Here you can focus on just the words, no fancy logos or buttons to distract you. Same as on the cleanup page, we have bookmarks. Easily add it to any row and you'll always remember where you left off or find the same row across all pages. To remove it, just hover over it and click x. Visit the main menu to change text size, hide timecodes and strip numbers or change the view order by switching between the chronological order as it was recorded, or as arranged to reflect the order you determined in the arrange tab.

Select is the heart of ScriptCut. What used to be a highlighter on a pdf is now a simple click. Click a line once to mark it green: the moments that are definitely in. Click a line twice to mark it yellow: your maybes. A third click clears it. Nothing here changes your transcript. You're choosing, not cutting. And that orange timer at the top is your live runtime: the exact length of everything you've selected so far. The footer breaks down the runtime of your must-haves and maybes. In the footer you can also select all, deselect all or just deselect the interviewer questions. As on the cleanup page, you can draw a select with your mouse for batch actions presented in this pop up. Whether a line is a keeper is sometimes determined by its delivery so just click the play button and listen to it. Don't love the whole line? Click and drag across just the words you want and they will be highlighted. You can do it with as many words as you like. If those word selects are a maybe, just click the whole row, and to deselect everything, click again. The running total updates as you trim, and every downstream step, playback, the PDF, your timeline, respects those word-level trims exactly.

If your goal was to just transcribe a file and clean it up, you can download your transcript straight from this page. Click Transcript and choose one of the many options. Our PDF's a nicely formatted and client ready. ScriptCut can create subtitles in all common broadcast standards in srt and vtt format. This turns ScriptCut into the perfect subtitle QC tool for professional use.

On ProAI Plan, one of the most time saving tools joins your toolkit. AI cleanup. It reads the transcript and tidies it for you. It identifies every sentence and every word and removes them for you. Think of off topic talk, slates, resets, everything that you'd remove before ever showing a client to make selects. When I say "every" I mean 95% because AI isn't perfect, but it gets pretty close. Before accepting the changes, scroll through the transcript and review. If you find a row that you'd like to keep, just hit the keep button. Then confirm the changes and your Transcript is reduced to only the content that matters.

ScriptCut will automatically create a new row for each sentence. However, sometimes a transcriber may misplace a period or misses when a sentence ends. To fix this, click into the line, put your cursor where the break belongs, and press Enter just like a regular text document. The line becomes two clean rows, and your runtime and word timing stay perfectly aligned. Splitting lines makes the script easier to skim on later, and it lets you keep one sentence and drop the next. It's the safe way to shape a long line without ever touching the words themselves. Merge with previous lets you merge two rows together.

Click Remove fillers and ScriptCut finds the hesitation sounds across your whole transcript, the ums, the uhs, the hmms. It shows you exactly what it found and waits for you to confirm before anything changes. Once you confirm, those words will get removed, and a small amber dot marks any row that has cuts. If the automation didn't catch a word, go to Edit Fillers and select any fillers you want to remove manually or click on the words you'd like to recover. Reset fillers in the row menu brings all of them back.

Cleanup reads like a document. Every line is editable text, with the speaker and timecode on the left. But there's one golden rule here: Don't delete or retype whole words. If you do, your text drifts out of sync. The goal here is to fix spelling mistakes, break up a sentences or merge them, and delete full off-topic rows. Don't just delete filler words, the next video will walk your through that process. If you do make changes that risk your sync health, a warning will appear. For batch operations, just hold down the mouse button and drag across all rows you want to select. A submenu will pop up presenting you with all available options. The speaker button let's you can change the speakers name and role as a batch. To change it per row, click the menu to the right, change speaker. You can assign a new speaker to the row or add a new one under custom. Once a change has been made on a row, a reset icon will appear that lets you reset all changes for that particular row. To search for specific words or sentences, click the search icon or Command F and type in your word. You can also easily replace a word for the entire transcript. Let's say ScriptCut was actually spelled ScriptCutter. Instead of replacing it one by one, just type it in here and click replace all. Done. Click the menu button in the top left, to change your text size, hide timecodes and strip numbers, or reset all edits. Before we wrap this up, one more little easter egg. When reading a book, you probably use a bookmark to easily continue where you left off. Transcript can run 90 or more pages, so we had to add this feature. Select bookmark on any row and next time you visit the page, it will automatically scroll to where you left off. The cool thing is, it will carry over to all other pages that follow, so you're always in sync. To remove it, either create a new bookmark or hover over it and hit the x button.

Last step: tell ScriptCut who's who. It lists the speakers it found and asks you to tag each one. Use Interviewer for whoever's asking the questions, usually off camera. Use Interviewee for your on-screen talent, whose name shows as a prefix in the transcript. And use Crew for anyone whose lines you want gone automatically. Tagging roles lets you clear the host's lines with a single click later, and it styles questions and answers correctly in your exports. Not sure who's who yet? Hit the little play button next to a name to hear a quick sample of that voice. When everything's tagged, press Start, and your transcript opens on the Cleanup page.

Attach the recording and you can hear, and see, every line while you or your client work within ScriptCut. You can upload MP3, WAV, or M4A audio, and MP4 or MOV video. It compresses the file automatically as it uploads, so it stays small and fast. If you're planning to make social clips, Keep the Video so ScriptCut holds the full picture. It must come from the same timeline as your transcript, and you can add or change it here any time. You don't need it to select and export, but hearing the pause before a punchline often turns a good pick into a great one.

Here's the usual grind: hours of footage, a scrubber, marked up PDF’s from clients. ScriptCut skips all of that. This is where a long, rambling recording becomes a tight, story-driven cut, before you ever open your editor. You work with words instead of a timeline. Read the transcript, tap the moments that matter, arrange them into a story while always keeping an eye on the actual runtime, and ScriptCut rebuilds every choice as a real timeline in your editing software. Podcasts, docs, vlogs, panels, talking heads, if it wasn't scripted, ScriptCut got it. Working with collaborators such as your client, directors, producers? Send them a review link to make their selects and approve before you touch a single edit. Come on in, let's look around.

Welcome home. Every project you make lives right here. Top right, three buttons do the heavy lifting: **New Folder** to organize by projects and clients, **Transcribe** to turn a raw recording into a transcript, and **New Project** to get rolling. Flip between grid and list, sort, or search by name. Each card wears little icons for what's inside: transcript, timeline, audio, and video. Click to open, or pop the three-dot menu to rename, duplicate, move, or delete, and don't sweat that delete, it just parks in the Archive for thirty days. The left sidebar rides along everywhere you go, and its tutorials are smart: you only ever see the ones that fit the page you're on. And see that chat bubble? That's your lifeline. Message your clients directly, more on that later, or tap **AI Assist**, ask it anything and get an answer on the spot. Next, let's start a project.

Click **New Project** and ScriptCut asks one thing first: how do you work? Go **Blank** and bring a transcript later. Pick **Transcribe** to hand over an audio or video file and let ScriptCut write the transcript for you. Or choose your editing software, and the whole ride, import steps, exports, even the tutorials, gets tuned to your editor. Itching to go social? The **Social clips** path turns long content into short, ready-to-post videos. Name it, drop in a cover image to make that share link pop, and hit **Create Project**. Done.

ScriptCut comes in three flavors. **Free** lets you select, arrange, and export an editor cut sheet, a great way to get your feet wet. **Starter** opens up the full exports: a formatted PDF, the timeline, subtitles, and spliced audio. And **ProAI** brings out the big guns, the AI tools, the publishing suite, your own branding, and a monthly pool of AI credits. Toggle monthly or annual, and annual quietly knocks off twenty percent. Transcription minutes and AI credits ride along with your plan, and you can top up any time. Prices shift now and then, so the live numbers always sit right here, in **Settings**, under **Plan**

You don’t need to transcribe and cleanup your transcript to re-use it. Drop the project that has that transcript into a folder, then create a new project inside the same folder and pick **Choose master transcript** instead of starting fresh. Every new project pulls from that one source, while you make your own picks and arrangement inside it. Fix a typo or a speaker name on the master, and every linked project updates on its own. No exporting, no re-importing, no copy-paste marathons.

Two housekeeping wins. Under **Settings**, **Storage** shows what each project is eating up, biggest first, with a bar if your plan has a cap. Tight on room? **Remove video, keep audio**, or **Remove audio** altogether, and your transcript and exports stay put, so the project keeps working. And under **App Integrations**, connect **Frame.io** to pull a video straight from your library, even a huge file, with no slow browser upload. Just pick it from the thumbnails and ScriptCut reels it in.

Experimenting is only fun when you can take it back. See that little version badge on a project, the V1? Give it a click. ScriptCut saves versions as you work, so you can try a bold recut, then rewind to any earlier point, like save states for your story. Landed on a cut you don't want anyone touching? Lock the version, and a client with the review link won’t be able to make any more changes . Here's the nice part: the moment your client approves, the project locks itself, right here, so nothing drifts after sign-off. Want to keep tweaking? Unlock any time. And for the full paper trail, Settings, History keeps automatic restore points too, a proper time machine for your edit.

o get a word-level transcript out of DaVinci Resolve, open the **Audio Transcription** panel, click the three-dot menu, and choose **Export Subtitles**. Now the settings that matter: set **Maximum** to **1 character per line**, set **Lines** to **Single**, and check **Include Speaker Name**. That one-character-per-line setting is the whole trick. It gives every word its own timecode, which is what lets ScriptCut cut to the exact frame. Save the file, then drop it into ScriptCut's transcript step." As of 2026, the transcription engine in Davinci Resolve isn’t perfect. Should you be unhappy with the results or experience timing issues, try to Transcribe within ScriptCut for more accurate results.

‍

Already have a transcript? Drag it onto the upload area, or click to browse. Upload either a SRTX, SRT, VTT or JSON subtitle file. To ensure accurate timing, ScriptCut needs a word level timecoded file. To check, open your subtitle file in a regular text editor and confirm that every word has a dedicated timecode range. As soon as the file is read, ScriptCut confirms it and tells you whether it found word-level timecodes and speakers. A green check means you're good to go.