Why Gigoff

Most AI video shows something. Gigoff explains it.

There are good tools for footage, presenters and templates. Gigoff is built for one job they weren't made for: making an idea clear, accurately, in an animated video.

Researched, not invented.

Gigoff reads around your question, or the paper, page or brochure you give it, and keeps every source with the script. You approve the facts before anything is made.

Drawn, not generated.

Every frame is built from vector shapes, maps, charts and real type, the way a designer would draw it. Words stay sharp and numbers stay exactly as the source has them.

Edited, not re-rolled.

Change one line, one colour or one scene and the rest stays exactly as you approved it. Every change is a version you can go back to.

Four ways to make a video

Same question. Four ways to answer it.

What a video tool builds its frames from decides what it's good at. Here is “How does a rate hike reach my EMI?” made each way.

Diffusion-based

Illustration

The frame is made of pixels predicted from noise, a new guess every time.

Good at cinematic scenes, moods and things that don’t exist.

Struggles with words, numbers and logos that must stay exact, and fixing one detail without a new take.

Curation-based

Illustration

The frame is made of stock clips and templates, matched to your script.

Good at quick social videos with a familiar look.

Struggles with showing the idea itself. The footage is about the topic, rarely the mechanism.

Diffusion + curation

Illustration

The frame is made of generated clips mixed with stock footage.

Good at variety, fast.

Struggles with both sets of problems at once: drifting text and generic footage, in two different looks.

Designed in vector

Gigoff

The frame is made of shapes, maps, charts and real type, laid out like a designer would and animated to the voice.

Good at explaining: how things work, what the numbers say, where things happen, step by step.

Not built for photoreal footage or people on camera.

Researched, not invented

From your source to the screen.

One number from our research-paper explainer, followed all the way through. Gigoff was given the PDF and asked to explain it.

01 · Your source
The paper says 4.8°C ± 1.2°C for doubled CO2.
Hansen et al., Global warming in the pipeline, Oxford Open Climate Change, 2023. Abstract, page 1.
“From ice age evidence, the authors estimate that doubling carbon dioxide would warm Earth by four point eight degrees Celsius.”
attachment-1.pdfdoi.org/10.1093/oxfclm/kgad008
02 · The script
Scene 2’s narration, with its sources kept beside it.
Every scene lists where its facts came from.
03 · You approve
You read the script and its sources first.
Nothing is designed, voiced or animated until you say go.
04 · On screen
4.8°C ± 1.2°C, exactly as the paper has it.
Real type and real numbers, drawn in vector. Never guessed into pixels.

Edited, not re-rolled

Change one scene. The rest stays put.

In our vertical “LLM or agent?” explainer, scene 3 put “LLM” in the loop where it should have said “Agent”. We changed that one scene. The other three stayed exactly as approved, and the first version is still there to go back to.

Scene 3 before: the loop says LLM — Scene 3 · before

Scene 3 after: the loop says Agent — Scene 3 · after

Untouched

Same voice, same timing, same frames. Only the scene you asked about was redrawn.

The usual way

Compared with how it’s done today.

Hiring a motion designer

A brief, a quote and rounds of feedback. Beautiful work, usually days or weeks for each explainer.

You write the brief and review drafts.

Doing it yourself

Templates and an editor. The research, layout and timing are yours, frame by frame. Hours for each one.

You do all of it.

Gigoff

Researched, written, designed, voiced and animated for you. A first cut usually in under half an hour.

You approve each step, and change any scene.

What designing in vector gives you.

Researched, with sources

Starts from a question, a link or a PDF, and keeps every source.

Exact words and numbers

Real type and real data, identical in every render.

Maps, charts and diagrams

The visual language explainers actually use.

Change one part

Fix a line or a scene; the rest stays as approved.

Both shapes, one story

Wide and vertical, laid out for each.

Voiceover and score

Seven languages including Hindi, or your own voice.

Your brand

Colours, fonts and logo on every explainer.

Every change a version

Go back to any earlier version.

Honestly

When Gigoff isn't the right tool.

We'd rather you pick the right tool than the wrong one twice.

You need real footage or real people. Gigoff draws; it doesn’t film or generate photoreal scenes.
You want a presenter on screen. An avatar tool, or your own camera, suits that better.
You’re editing your own recordings. Use a video editor, and drop Gigoff’s explainers in where they help.
You’re making a cinematic brand film. That’s still a job for a production team.

Questions when you’re comparing.

Is it AI-generated?

AI does the research, writing and design. But every frame is built from vector shapes and real type, not generated pixels, so words and numbers stay exactly as written.

Can I check the facts before it’s made?

Yes. You read the script with its sources and approve it first. Nothing is designed, voiced or animated until you do.

What if one number is wrong?

Change that scene only. The rest of the video stays as you approved it, and every earlier version is kept.

Can I drop it into my own edit?

Yes. Export wide or vertical, with the voice on or off. For a transparent background, export WebM or MOV (ProRes 4444).

Who owns the video?

You do, apart from third-party material such as licensed music. We don’t use your content to train AI models.

See the difference on your own idea.

Your first explainer is free. No card needed.

Make an explainer See examples