Reactions that match the line
Every sentence is read for the emotion behind it and given an avatar pose to match, animated with a preset that does not repeat itself two lines running.
Desktop Windows, macOS and Linux — one app, one purchase
Flomoji reads what you said and builds the edit around it: a reacting avatar, b-roll, sound effects and video effects, placed on the beat. You get a timeline you can change and an MP4 you can post — rendered on your own machine.
Two free videos every week, no card needed No watermark on anything you export Your footage never leaves your machine to render
No timeline to build by hand, no scrubbing for reaction frames, no hunting for a sound effect that lands on the right syllable.
Any talking-head clip, vertical or landscape. It is transcribed word by word, so every later decision is pinned to the exact moment you said it.
Each sentence is read for emotion and given a matching avatar reaction, with b-roll over the visual moments and sound effects on the beats that deserve one.
Move, trim or delete any clip, drag in your own, add text on the canvas. Then render to MP4 — on your own machine, at full quality, with no watermark.
Most tools in this category place an effect wherever they can. The thing that separates a good edit from a busy one is knowing what to leave out, so the brain reads the whole video first — what each sentence is for, and what the video is as a whole — and then decides what survives together.
A desktop tool that treats your footage, your machine and your time as yours.
Every sentence is read for the emotion behind it and given an avatar pose to match, animated with a preset that does not repeat itself two lines running.
Stock photos and video are placed over the lines that describe something, covering roughly half the runtime in clips of a few seconds. Stills get a slow Ken Burns move; video clips do not, because the footage already moves.
Cues are snapped to real word boundaries, loudness-matched so none of them jumps out, and your voice ducks underneath instead of fighting them.
Select, move, trim or delete any clip on the timeline, drag in your own images and sounds, and lay text over the canvas. Then export again from where you are.
The layout profile follows your footage, and the bottom quarter of the frame is always left clear — so burning captions in later never covers anything.
New avatar sets and sound packs arrive without a reinstall, and the app updates itself in the background. Every update is verified before it is installed.
No upload step, no shared render queue behind other people's jobs, no per-minute export bill. Your own hardware is the only thing deciding how fast it finishes.
Your licence is checked on your own machine against a signed token, with a fortnight of grace built in. Our servers going down — or going away — does not stop you working.
Nothing is stamped on your export, on any tier. No second installer, no “recommended offers”, no upsell banner inside the app you paid for.
Three things a good editor does by hand, on the beat, and then stops doing before it gets tiring. Each one is placed from what you actually said — and each one can be dragged, retimed or deleted afterwards.
“…and everything just froze.”
One word off the line you are speaking, thrown on screen as it lands. It picks the word that carries the sentence — never a filler word, never a number you did not say — and holds it for under a second.
1.18× punch in
“…and that is the part nobody posts.”
A slow push in on the lines you get animated about, eased in and out so it reads as camera movement rather than a jump. It never zooms twice inside six seconds, and never over more than a third of the video.
Effect on
Untouched
“…and then the room went quiet.”
Eight grades, six texture treatments and six impacts, each with a strength you set. Drag one onto the effects row yourself, or leave it to the brain, which will usually choose one grade and one flash for a whole video.
A callout is lifted from your own transcript, pinned to the moment you said it. Nothing is written for you and nothing is invented.
Hard ceilings, not suggestions: two devices at once, one loud effect per video, four impacts in total. A busy edit is a worse edit.
Graphics, zooms and effects are rendered locally by ffmpeg. No extra API key, no per-effect charge, nothing uploaded to draw them.
Every automatic choice lands on the timeline as an ordinary clip. Move it, trim it, change its strength, or delete it and put your own there.
Leave every one of these switched off and you still get the avatar, the b-roll and the sound. Nothing here is required to get a finished video.
Everything the app decided is sitting on a row you can reach. Nothing is baked in until you export, and no decision is hidden behind a “regenerate” button.
Six rows, each one draggable and trimmable. The striped blocks on the Video row are ranges you cut — click one to put it back.
Renders fail. Machines sleep. A four-minute export gets interrupted at exactly the wrong moment. Most of the engineering here went into what happens next.
At a glance
Browser-based editors are convenient right up to the moment the render matters.
| Web-based editors | Flomoji | |
|---|---|---|
| Where your footage goes | Uploaded to a stranger's server | Stays on your disk |
| Export length limits | Almost always | Your free space is the limit |
| Watermark on export | Until you upgrade | Never, on any tier |
| Interrupted renders | Start again | Picked up on the next launch |
| Render queue | Behind everyone else's jobs | Your machine, immediately |
| Works offline | No | Everything but the first transcription |
| What it costs | A subscription, and only a subscription | Buy it outright, or subscribe, or neither |
Flomoji is built for turning the recording you already made into something worth posting.
Post daily without spending an evening on each cut. Record once, let the reactions and b-roll go in on their own, adjust what you disagree with.
Turn a talked-through explanation into a lesson that holds attention, without learning an editing suite to do it. Tutorials get no grade and fewer cutaways, because the picture is the point.
Take the good three minutes out of an hour and give it the visual pacing a feed actually rewards, in one pass rather than twenty.
Get a week of posts out of one sitting in front of the camera, without booking an editor or waiting on someone else's turnaround.
Feed in a half-hour recording, let it find the moments worth keeping, and cut each one down to something a feed will actually watch.
Client work under NDA, internal training, anything sensitive. Rendering happens on your own machine, so the video never leaves it to be made.
Editing tools in this category are almost always a monthly bill that never stops. You can pay monthly here too — but you do not have to, and that is the point.
Monthly
$9/month
Lifetime
$67once
$67 is what seven and a half months of the monthly plan costs. Keep using it for three years and paying monthly would have cost $324 — the lifetime licence is $257 less, and it does not care how long you keep it.
The free tier is a real product, not a demo: two videos every week, forever, at full quality and with no watermark. A licence raises that to 30 videos a month at thirty minutes each, and unlocks the post-render editor.
Two videos a week, forever.
$0.00 one-time
Thirty videos a month.
$9.00 per month
Pay once. Thirty videos a month, forever.
$67.00 one-time
Content you hold the rights to: your own uploads and recordings, Creative Commons and public-domain work, and anything you have been given permission to keep offline. You are responsible for the rights to what you save, and our licence agreement says so plainly rather than burying it. If you are not sure you may save something, the answer is that you should not until you are.
No. The free tier renders at the same quality as a paid licence — no watermark, no resolution cap, no nagging overlay. What it has is a ceiling: two videos a week — forever, not for a fortnight — up to five minutes each, and no post-render editor. A licence raises that to 30 videos a month at thirty minutes each and opens the editor.
Only for the first pass over a new recording — transcription and planning run online. Rendering does not. Your licence is verified on your machine against a cryptographically signed token, with a 14-day grace window if we cannot be reached at all. A licence server that can strand a paying customer is a badly designed licence server.
Only if you want it to be. $9 a month bills automatically and you can cancel it whenever you like. $67 once charges you a single time and never again — no stored card, no renewal date, no price rise in two years. Most comparable tools only sell you the first kind.
You drop to the free tier. The app keeps working, your files stay where they are, and nothing is held hostage. When you renew, it extends the licence you already hold — same key, same activated devices, nothing to paste in again.
Yes, and without asking us. Release the old device from your account, activate the new one, and carry on. Each plan includes a number of self-service moves per 30 days.
Your footage is not uploaded for rendering — that happens on your own machine. Audio is sent for transcription so the timeline can be built, and nothing else about the file is kept. Licensing requests carry your key and a device identifier, because that is all the API accepts. The privacy policy lists it exactly.
The installer contains the application and nothing else: no bundled toolbars, no second installer, no browser extension, no advertising SDK. Components the app fetches after installation are verified against a published checksum before they are used, and a component that fails to run is rolled back automatically.
Try it before you decide anything
Install it, save something, and see whether it behaves. Buy a licence only once it has earned one.