Nalo Seed video automation

You already shot the video. Getting a week of clips out of it should not take a week.

We run a pipeline that takes one long recording, finds the moments worth posting, and cuts them into vertical clips with captions burned in and your brand on them. The part that usually goes wrong, the framing, gets looked at by a person before anything renders.

Talk to us about your footage See how it works

How it works

It reads the whole thing

The video is transcribed, and the transcript is used to pick the moments worth cutting. You get windows with a start, an end, and the words inside them, not arbitrary sixty second slices.

It draws the framing before it renders

For every chosen clip it produces a framing sheet: the full frame, with the vertical crop outlined on top and everything outside it dimmed. That takes about ten seconds a clip.

A person checks it

You look at the sheets and nudge the ones that are wrong. This is the step most tools skip, and it is the reason clips come back with the top of someone's head missing.

Then it renders

1080 by 1920 MP4s, captions burned in, in the client's own brand colours with their logo. Five to eight minutes of machine time per clip, and no one is watching it happen.

The hard part, drawn to scale

A vertical clip cut from a normal widescreen recording keeps 31.6% of the width. Two thirds of the picture is thrown away. If the speaker is not standing in the part you kept, you get a clip of their shoulder.

This is a diagram of the mechanism, not a still from your footage. On the first real source we ran, ten sampled moments gave six usable framings, three with a face clipped at the edge, and one where the speaker was outside the crop altogether. Catching that on a framing sheet takes about ten seconds. Catching it by rendering takes five to eight minutes.

Measured, not estimated

About 10 seconds: to draw the framing sheet for one clip
5 to 8 minutes: to render one 60 second clip
About 40x: cheaper to look than to render

Taken from a sixteen minute 4K source on a laptop. Your numbers will move with the footage.

What it does today, and what it does not

This is a working internal tool that we run for our own clients. It is not a product you log into, and we would rather tell you where the edges are now than after you have sent us a file.

It does

Transcribe a long video and choose the moments to cut
Draw a framing sheet for every clip before rendering
Burn in captions timed to the words
Use your brand colours and your logo
Output 1080 by 1920 MP4s ready to upload

It does not

Post anything to any platform. A person uploads the clips.
Add trending audio. The licence on those sounds does not cover commercial use off-platform, so we do not touch it.
Follow a speaker who moves. A wide two person shot still needs a human call.
Give you a dashboard. Today you talk to us and we run it.

Want your footage cut this way?

Tell us what you shoot. If it is a fit we will cut a couple of clips from something you already have, so you can judge it on your own video rather than on a landing page.

Looking for the AI video generator instead? That is over here.

🎬

✨

🔥

🚀

🎥

LOADING...

Getting your vibe ready! 🔥