In development

One recording. A month of LinkedIn video.

Record the call you were going to have anyway. Clips finds the passages worth posting, cuts them so they start and end on a whole thought, captions them, and puts them inside your brand frame.

Try it with one recording See how it works

Not open yet. The pipeline is being built and tested. Leave an address on the try page and you will hear when a slot opens.

How it works

Three steps, and you only do the first one

The work is in the selection. A thirty minute conversation holds maybe eight passages that stand on their own, and finding them is the part that otherwise eats an afternoon.

Step one

Upload the recording

An interview, a panel, a webinar, a customer call, a two-person video call. Whatever you already recorded.

Step two

It reads it and chooses

Transcribed to the word, so cuts land on whole sentences. Passages that make a point on their own become candidate clips. Where two related thoughts sit minutes apart, they get stitched.

Step three

You get finished files

Cropped to the ratio you chose, captioned inside the safe area, framed in your card. Download and post, or send them to whoever runs the page.

The frame

A frame is what makes it look posted on purpose

A raw crop of a video call looks like a raw crop of a video call. The same clip inside a card with your colours, the speaker's name and captions across it reads as something the company published.

The frame is optional, and a bare clip is a perfectly good output. If you want one, there are two ways to get it.

You already have a cardUpload the PNG your designer made. The window is read out of the artwork itself, so whatever shape it is, the video takes that shape exactly. Corner radii that differ from each other survive, because nothing assumes a rounded rectangle.
You do not have a cardPick two background colours, a stroke colour and a logo, and choose one of three layouts. It builds the card and every clip you ever make uses it.
Optional motionThe picture can open like an old television switching on, and a small light can travel slowly around the frame's edge. Both are switches, off by default for brands they do not suit.

Speaker name

Role, company

CAPTIONS LIKE THIS

YOUR LOGO

Layout, colours, corners and type all come from your brand, not from a template you have to live with.

What you get back

Files, not a dashboard

Every clip comes back as an MP4 you own, in the ratio you asked for. There is nothing to log into afterwards and nothing that stops working if you cancel.

Ratios

4:5 and 1:1

4:5 is the default. It fills a phone in the LinkedIn feed without going full screen, so your post copy and the comment count stay visible next to the video. 1:1 is the safe alternative. Vertical is only offered when the footage was shot vertical, because a 9:16 crop out of a 16:9 recording throws away most of the frame and softens what is left.

Captions

Burned in, inside the safe area

Most feed video is watched with the sound off. Captions are burned into the picture rather than uploaded as a sidecar, sized to the frame, and kept clear of the platform's own overlays at the bottom and right. The highlight colour is yours to pick.

Subscribers

Your brand, remembered

An app.ional.ai subscriber has brand guidelines stored already, so the card is generated from them and every clip matches everything else the company puts out. Without a subscription you upload a logo and pick colours each time, which works and takes a minute.

Subscribers

Part of the wider set

Clips runs on its own here. Inside app.ional.ai it sits alongside the other skills, so a recording can feed the content calendar and the clips can be scheduled rather than downloaded.

Straight answers

What it does not do

Worth knowing before you upload something.

Current limits

It cannot rescue bad audio. Selection runs off the transcript, so a recording with heavy crosstalk, a room echo or a bad line produces poor cuts. Phone or headset audio is fine; a laptop microphone across a meeting room is not.
It will not invent a clip that is not there. A conversation that never lands a self-contained point yields nothing worth posting, and it says so rather than shipping filler. Being asked a direct question tends to produce the best passages.
Vertical from horizontal is refused. A 9:16 crop from a 16:9 recording keeps about a third of the width and needs roughly a two times upscale, which looks soft. Shoot vertical if vertical is what you need.
It is built for the talking head. A person speaking to camera, framed tidily, captioned. Not a fast-cut edit, not screen recordings, not music-led video.
Two-person calls work, with one caveat. A recording where the camera follows whoever is speaking is fine, and an exchange between two people can become a clip. What is not ready is the card: it names one speaker, so a clip carrying both currently needs a card with no name on it. Grid views where both faces sit side by side do not crop well to portrait.
Nothing posts on your behalf. You get files. Publishing stays with you.
Anyone on the recording is on the video. Get the agreement of the people in it before you upload, and again before you publish. This is a small industry.