How can we help?

Sign in with your Whop account on the app page, pick a video file from your device (or paste a direct link to one), choose how many clips and how long, and press Find moments. Review and trim the suggestions, then export the ones you want as MP4 files. The app takes you through this in five steps: see “How the steps work” below.

The app page has five steps along the top: Video, Captions, Clips, Style and extras, and Export. Only the current step is shown, with a Back and a Next button at the bottom, and you can click any step that is open. Nothing is lost when you move between them: your transcript edits, clip titles, Fix words list, Tighten pauses settings and Brand kit stay as they are, and a download or export that is running keeps running while you look at another step.

1. Video: choose a file or paste a direct link, name the project and see what your plan allows. When the video has loaded you go to Captions.
2. Captions: generate them automatically on your device, upload an .srt or .vtt, type your own, or choose none. You can skip this step. Fix words lives here.
3. Clips: set how many and how long, press Find moments, then edit titles, start and end times, and untick clips you do not want.
4. Style and extras: caption style (including Word highlight), framing, and for Pro Follow the speaker, Tighten pauses and the Brand kit. Each is a card you can open; the card says in one line whether it is on.
5. Export: a summary of what will be exported, Export all or one clip at a time, the downloads, the zip with the posting pack.

A step opens when it can be used: Captions and Clips need a video, Style and extras needs clips, and Export needs at least one ticked clip. A step that is not open yet says why. The small dot on a step means something is still working in the background, for example captions or an export. The address shows the step (for example /app#clips), but nothing is saved: if you reload, you start again on Video.

Any video your browser can play. MP4 and WebM work in most browsers; other formats such as MOV depend on your browser. If the page cannot read the file you will see a message.

No. Reelify can’t download from YouTube, Vimeo, TikTok, Instagram, Spotify, Apple Podcasts, Twitch, Facebook, X and similar sites, and a link to one of their pages shows a short explanation instead of being fetched. If it is your own video, download it from your creator dashboard (YouTube Studio lets you download your own uploads) and drop the file in the app.

A direct link to one video file: an https address that ends in the file itself, such as https://example.com/video.mp4 (.mp4, .webm, .mov, .m4v or .mkv). Your browser downloads the file straight from that site into this tab, so Reelify never sees the link’s contents, and the file then goes through the same limits as a file you pick: Free videos up to 30 minutes, Pro up to 3 hours. Link downloads are limited to 500 MB on Free and 2 GB on Pro, because the file is held in your browser’s memory. A link has to be public: no sign-in, no username or password in the address, and nothing on a private network or this computer.

The usual reasons: the site does not allow other websites to download its videos (browsers block that, and the only fix is to download the file from the site yourself and drop it in the app); the link opens a web page rather than a file; it needs a login or has expired (403 or 404); it is an audio file (Reelify makes vertical video clips, so it needs a picture, and audio-only files are not accepted); it is an http:// link (only https works); the file is over your plan’s link limit; or it is a streaming playlist (.m3u8 or .mpd) rather than one file. Each of these shows its own message. Cancelling a download is fine: you will see “Download cancelled.” and nothing is kept.

If you have captions (automatic ones, or an uploaded .srt or .vtt, including your edits), Reelify ranks stretches of whole sentences, about 20 to 60 seconds and never over 90, using simple signals: how the first line opens (a question, “here’s”, “the truth”, “nobody”, a number), whether the clip starts and ends on complete sentences, whether it starts without leaning on an earlier sentence (not “and”, “but”, “so” or a dangling “it”), how loud it is, how steadily people talk, and whether it ends on a sentence followed by a pause. Cuts move into the pauses around the sentences, so no word is cut. Each moment lists only the signals that actually applied. Moments are shown in time order. If a transcript has no punctuation, Reelify follows pauses and caption lines instead and never calls a clip a complete thought. With no captions it falls back to ranking by loudness only, with a limited sample of picture changes, and tells you so. You can switch between the two, and use “Re-rank from the current transcript” after editing the text. These are simple rules, not AI summarising: Reelify does not understand what is said, so always review and adjust.

A title is the first sentence of the clip, taken from the captions and shortened. Nothing is written for you, and automatic captions can contain mistakes, so every title is editable per clip.

No. Your video is read and cut inside your browser using ffmpeg compiled to WebAssembly, served from this site. Reelify does not upload, store or watch your video, and it keeps no history of your clips. Follow the speaker, a Pro option, looks at frames of your video on your device only; the single thing it downloads is a small face-detection model file from huggingface.co. If you paste a direct link to a video file, your browser downloads it straight from the site that hosts it; the download does not go through Reelify.

It runs on your own device, so it depends on your computer, the video size and the clip length. Finding moments in a long video can take several minutes, and a clip exported with Tighten pauses takes longer than a normal one. On our test machine a 1-minute clip took roughly a minute to export; your device may be faster or slower.

Yes, automatically, on your own device, for English speech. In the app choose Captions, then Automatic, then Generate captions automatically. Reelify listens to the video’s audio with an open Whisper English model (you choose the Caption quality: Fast, Better or Best), writes the words with their timing, and puts the text in a box you can edit before exporting. All three qualities are available on Free and Pro, with Classic captions on Free (the other 5 styles are Pro). You can still supply your own .srt or .vtt file, or type caption text for each clip by hand.

The first time you use a caption quality, your browser downloads that model from huggingface.co and its download network (hf.co), a third-party host: about 42 MB for Fast, 78 MB for Better or 250 MB for Best (about 63, 111 and 323 MB on devices that use their graphics card). It also loads about 7 MB of engine from this site. Each model is kept in your browser on its own, so switching back to one you have used does not download it again, and later runs work offline. Only those model files are fetched: your video, audio and transcript are never sent anywhere. You can cancel the download at any time.

None of them is perfect, so read the transcript before you export. Bigger models make fewer mistakes on average, but are a bigger download and slower: Fast (Whisper tiny, about 42 MB) took about 0.2 times the video’s length in our test and about 0.8 GB of browser memory; Better (Whisper base, about 78 MB) took 0.3 to 0.5 times the length and 1.0 to 1.2 GB; Best (Whisper small, about 250 MB) took about 1 times the length (a 10-minute video takes about 10 minutes) and 1.5 to 1.7 GB. On our fastest and noisiest test recordings about 1 word in 4 or 5 was wrong with Fast, about 1 in 7 or 8 with Better and about 1 in 13 with Best. On one slower voice, Fast and Better each left out a stretch of speech (roughly 3 in 10 words wrong) while Best did not (about 1 in 10). Those figures come from computer-generated test speech on one test computer, not from real people talking, so real results vary with the voice, accent, microphone, background noise and music, and your device. Start with Fast for clear speech; pick Better or Best for names, numbers, fast talkers or noisy recordings if your device can take it, and use Fix words and the transcript box for what is left. Speech in other languages is not supported. If the browser runs out of memory, close other tabs, pick a smaller quality or use a shorter video. Switching quality changes nothing until you press Generate captions again.

Fix words is a list in the Captions card for words the recognition keeps getting wrong, such as names and brand words. Type what was heard (for example “wap”) and what it should be (“Whop”), then Add. It matches whole words whatever the capitals, and carries the capitals over: ALL CAPS stays all caps and a Capitalised word stays capitalised. A phrase can be several words either way (“e commerce” to “ecommerce”). New captions use your list as they are made; for captions already on screen press Apply. The change goes through the transcript, the timed words behind Word highlight and Tighten pauses (every word keeps its place in time), the moment titles and so the posting pack, file names and exported captions. It does not re-run the recognition and it cannot find mistakes for you. Under the transcript, Find and replace changes any text with a count, and Undo puts back the last change. You can import or export the list as plain text, one pair per line: heard => should be. Up to 200 pairs.

In this browser, on this device, in its local storage. The list is not uploaded to Reelify or anywhere else, and it is not shared with your other devices or browsers. Clearing the site’s data (or using a private window) removes it, so export the list as a text file if you want to keep it. Fast, Better or Best is remembered the same way.

Most editors and caption tools can export .srt or .vtt, and YouTube Studio lets you download subtitles for your own videos. Reelify only reads the file you give it.

There are 6 styles. Classic is on every plan; the other 5 (Bold, Highlight, Word highlight, Minimal and Boxed) are Pro. Word highlight shows 2 to 4 words at a time and lights the word being spoken in the accent colour, a little larger, burned into the exported video. It needs word timing: with automatic captions the timing is exact. With an uploaded .srt or .vtt, or typed text, only the line timing is known, so the words in each line are spread evenly and the app labels it “Word timing: approximate”.

It shortens the silences between words in the clips you export, so slow stretches move along. Where it cuts, the two sides are joined with a very short crossfade (5 to 15 milliseconds) so there is no click, and the quiet room sound is kept so the pause does not turn into dead silence. It does not speed anything up or slow anything down, does not remove words, does not remove filler words such as “um” or “uh”, and does not clean up noise. Breaths, laughs and other sounds between words are kept unless you turn on “Trim loud breaths”. Pauses that look intentional (a very long dramatic beat, a pause before a number or a shouted word) and anything it is unsure about are left alone, and the app lists them with the reason.

A preset sets how much pause is left behind and how long a pause has to be before it is treated as a long one. Short-form trims the most, then Voiceover and Conversation (the default); Interview, Educational and Storytelling leave a little more room; Documentary leaves almost everything alone and only shortens very long silences. Reelify does not detect what kind of video yours is, and it does not understand what is said: you pick the preset that fits, and changing it only changes the numbers.

How much pause is left behind after a cut. Light leaves the most (the smallest change), Tight leaves the least. Words are never cut at any strength: a pause is only shortened where the audio itself is quiet, with a little padding kept on each side of the neighbouring words.

Yes. For the selected clip the card shows the original length, the tightened length, the time saved, the number of cuts and how many automatic checks passed, then lists every proposed cut with where it is, the original gap, the new gap, the kind of pause, a risk level and the reason. Each cut has its own switch, and the summary updates straight away. “Reset to suggestions” puts them all back. Changing the clip’s start or end, the transcript, the preset, the strength or the breaths option recalculates the cuts; a switch you turned off is kept if the same cut comes back. The setting applies to every clip you export from the project.

To know where each word starts and ends. Automatic captions come with the time of every word; a .srt or .vtt file, or text you type per clip, only has the time of each line, which is not enough to find the pauses between words safely. If you have only those, the card is switched off and explains why, with a button to generate captions. Edits you make to the automatic transcript are used, and the timing follows them.

It is part of Pro along with longer videos, more clips, all caption styles and no watermark. Free can see the card and what it does; turning it on needs Pro. The check is made on the server before an export, like the caption styles.

No. After each tightened export Reelify measures the file it made: the length against the plan, that sound and picture end together, that no dead-silent gap was added, that there is no jump at each join, and that the time spent speaking is kept. Those are numbers. They cannot tell you whether it sounds natural, whether a pause was part of the joke, or whether a cut feels wrong. Automatic checks can’t replace listening: listen to the result before posting.

The clip is exported without tightening, and the clip’s card says so, for example “The pause plan was not used (…), so the clip was exported without tightening.” The same happens if the finished file fails one of its own checks. Nothing is lost: you get the normal clip.

Automatic captions are English only, and the pause types lean on English cues (a question mark, and words such as “so”, “now”, “anyway” or “however” that often start a new topic), so it is oriented to English speech. Other languages are not supported.

Yes. Reelify reads the clip’s audio once to find the pauses, then cuts and joins the picture and the sound and measures the result, all on your device, so a tightened export takes longer than a normal one, and longer clips and slower devices take longer still.

A wide video has to be cropped to a vertical 9:16 window. By default you pick one fixed position (left, centre or right). With Follow the speaker, Reelify looks at about one frame every 0.75 seconds of each clip, finds the largest face, and moves the window sideways to keep it in view, smoothly and without sudden jumps. It only pans: it never zooms, and it does not touch the sound, the speed or the captions. If no face shows for a stretch, the window stays where it was; if none is found in the whole clip, it stays in the centre. A video that is already vertical, or nearly, has no room to pan, so it is exported as it is and the app says so. It works together with Tighten pauses: each kept part of a clip is cropped from the speaker’s position at that moment.

Not perfect. It finds faces, not speakers: it does not listen, so it cannot tell who is talking. It can miss someone in side profile, a small or far-away face, someone looking down or partly hidden, poor light, and it can pick the wrong person when several are on screen (it follows the largest face and only moves to someone else when they are clearly bigger or stay in view). Under the card you see sampled frames of the clip with the 9:16 window placed on each and the face it followed marked, plus plain notes (for example, how often no face was found). Check that preview before you export. For any clip that looks wrong, choose “Use manual position instead”.

The first time you use it, your browser downloads one small face-detection model file, YuNet (about 0.2 MB, MIT licence), from huggingface.co, a third-party host. It also loads the same WebAssembly engine from this site that automatic captions use (about 7 MB, kept in your browser). The file is checked against a fixed fingerprint before it is used and then kept in your browser, so later runs do not download it. The faces are found on your device, in a background worker: the video frames are never sent to Reelify, to Hugging Face or anywhere else, and nothing is saved on a server. You can cancel the download or the analysis at any time. If the model cannot be downloaded (you are offline, or huggingface.co is blocked) or this device cannot run it, the app says so and uses the manual position instead.

It is part of Pro with longer videos, more clips, all caption styles, Tighten pauses and no watermark. Free can see the card and what it does; switching it on needs Pro. The check is made on the server before an export, like Tighten pauses.

A little. Looking for the speaker takes some seconds per clip the first time (frames already looked at are reused if you move a clip’s start or end), and the export itself is cropped frame by frame instead of with one fixed window.

A Pro feature that puts your look on every clip you export: your logo in a corner (size, opacity and margin you choose), your caption colours (text, highlight and outline or box), one of six fonts, where the captions sit (lower third, middle or upper) and how big they are. It works with every caption style, including Word highlight (the highlight colour is the lit word), and together with Tighten pauses, Follow the speaker and batch export. Presets: “Reelify default” and your saved kit, and “Reset to default” puts everything back. You can also export the kit as a small JSON file and import it in another browser. Free cannot use it and keeps the Reelify watermark; Pro exports have no watermark. A Free request that asks for it anyway is refused.

Only in this browser, on this device (in its IndexedDB storage; if the browser does not allow that, in memory for this tab only, and the card says so). Your logo is read, checked and redrawn in your browser: it is not uploaded, and Reelify’s server never sees it. The kit file you export is saved by your browser to your own computer. Clearing the site’s data removes the saved kit, and so does “Reset to default”. The server only learns that a Pro export asked to use the brand kit, so it can check your plan.

PNG, JPEG, WebP or SVG, up to 2 MB and 2048 pixels on the longest side (at least 16 pixels on each side). The file is really decoded to check it is an image, then stored as a PNG of at most 512 pixels on the longest side, so a huge file is not carried around. Transparency is kept. An SVG needs a width and height or a viewBox. Files that are the wrong type, too large, too small or damaged are refused with a message saying why. If a saved logo cannot be read when you export, that clip is exported without the logo and its card says so; the export is not lost.

You choose a corner, a size (a share of the frame width, keeping the logo’s shape) and a margin. With “Keep inside the safe zones” on (the default), the logo also stays clear of the strips at the top, bottom and sides of the vertical frame that many apps cover with buttons, names and their own text. Apps differ, so this is a cautious guide, not the rule of any one app. The preview shows the guide, and the card warns you if the logo would sit on top of the captions.

Reelify measures the contrast between your text colour and the outline colour, between the highlight and the outline, and checks the highlight is clearly different from the text (WCAG contrast ratios: 4.5 or more is comfortable, 3 or more is acceptable for large text). It warns you when a pair is hard to read; it does not stop you, because the colours are your choice. The measure is a guide: a busy picture behind the captions can still make them hard to read, so look at the preview, and at the “real frame” preview, which draws the first caption line over a frame of one of your clips.

Inter, Montserrat, Poppins, Oswald, Bebas Neue and Anton, plus the Reelify default. All six are open fonts under the SIL Open Font License 1.1. The font files are bundled with Reelify and served from this site: nothing is fetched from a font service while you use them. The licence text and each copyright line are at /fonts/LICENSES.txt and in the “Font licences” note on the brand kit card. The bundled files cover Latin letters; other scripts are drawn with your system font. If a font cannot be loaded, the captions use Reelify’s default font and the clip’s card says so.

Yes, on Free and Pro. “Export all selected clips” makes every clip you have ticked, one after another, with overall and per-clip progress. You can cancel at any time: clips that already finished are kept. If one clip fails you see the error for that clip and the others carry on (a refusal from your plan stops the batch, and says why). Then “Download all (.zip)” in Your exports puts the finished clips and the posting pack in one zip file. The zip is built in your browser and nothing is uploaded; the clips are stored in it as they are, without compressing them again, so it is quick. Free clips keep the watermark and Free’s clip limit still applies. A very large set of clips (over about 600 MB) can use a lot of memory, so the page warns you and you can download the clips one by one instead; above 4 GB a single zip is not possible.

For each clip it is a few lines of draft text for when you post it: the title (the same editable title that names the file), a short description, a hook line, a few suggested hashtags, and the clip’s timing in your video. It is simple text taken from the clip’s own captions, not AI, and nothing is invented. The description is the first one or two sentences of the clip (shortened with an ellipsis if long), the hook is its first sentence, and the hashtags are the most frequent meaningful words in the clip (common words, numbers and very short words left out), so they are suggestions drawn from the words in the clip, not from what is trending. Automatic captions can contain speech-model mistakes, and those carry into the text, so read it and edit every field before you post; shorten it to fit where you post. Edits are kept for this visit only. Without captions the pack has only the title and timing, and the app offers to generate captions. The zip holds posting-pack.txt, posting-pack.csv (opens in a spreadsheet) and posting-pack.json.

Free: videos up to 30 minutes, 5 clips per video, Classic captions (1 of 6 styles) and a Reelify watermark. Pro ($9 a month): videos up to 3 hours, up to 15 clips per video, all 6 caption styles (5 are Pro), no watermark, Tighten pauses, Follow the speaker and the Brand kit. The app shows your limits and how many clips you have used before you start.

When you sign in, the server checks your active Whop membership to Reelify Pro. If the check fails, Free limits apply until it succeeds.

Upgrade from the pricing page; Pro is billed by Whop. You manage or cancel it from your Whop account. Refunds and cancellations follow Whop’s checkout terms.

Automatic captions in languages other than English, downloading from YouTube, podcast or other page links (only direct links to a video file work), saved projects and history, and sharing links. Tighten pauses shortens silences only: removing filler words, breaths or noise is not something Reelify does. Follow the speaker follows faces, not voices: it can miss a speaker or follow the wrong person, and it never zooms or reframes to a different shape. The brand kit has one logo, one set of caption colours and one font per export (no per-clip kits, no custom font upload, no animated logo). The clip finder is basic and the export is one size (720×1280).

Exports live in your browser tab only. Download each clip (or the whole batch as a zip) when it finishes; refreshing or closing the tab clears them, and any posting-pack edits with them, because nothing is saved on a server.

Yes, by someone determined. Before an export the server checks your plan against the video length, clip count, caption style and whether you asked to tighten pauses, follow the speaker or use the brand kit, and hands back a short-lived signed permission (about 20 minutes) that the export must present; Free exports always carry the watermark. But all processing is client-side, in your browser, so the export code runs on your device and a determined person could modify it to skip these checks. They are there to keep Free and Pro fair, not to stop a skilled user, so please respect them. If the server restarts while you are exporting, the page asks for a fresh permission once and carries on.