Ways in
Five ways to get material into a project. They converge on the same thing — a project with scenes — so the rest of the docs applies whichever you pick.
Record a browser tab
The main path. The decodx extension captures a tab, tracking cursor and clicks so scenes can be cut on your actions and zooms can follow them.
Use it when the thing you're explaining lives in a browser and you can perform it live.
Upload an existing video
Bring a screen recording you already have — from Loom, Zoom, QuickTime, or anywhere else. decodx segments it and drafts narration the same way.
Upload video opens a short form. Add your clip first. Auto narrate is picked for you: it transcribes what you said in the clip and re-voices it cleanly. Choose Write my own if you'd rather type the script. Pick a voice and press ▶ to hear it. Templates, intro and outro titles, speed, size, music and Also create a written guide are under More options. Your draft saves as you go, even before the upload finishes. The line next to Create video shows how long the video will be and how many AI minutes it uses.
Use it when the recording already exists, or when what you need to capture isn't a browser tab.
Start from a prompt
Describe the video you want. decodx researches it, plans the story and the scenes, and makes the video.
In the app, the prompt box on Home and + New video in the sidebar both open one start screen with a tab for every way in. On Describe it, write what the video is about, then choose Make my video. decodx reads your Brand Kit (switch Use my Brand Kit off to leave it out), opens any links in your prompt, and searches the web for facts and figures about the topic, each kept with its source. It picks how to tell the story, writes every scene (animated motion scenes included), checks the length, your brand rules and that every number has a source, then builds the video. You watch each step as it happens, and land on the finished video with What we used (the Brand Kit, the sources) and Show the story. Drafting is free; the render uses AI minutes, and the estimate sits next to the button. Choose a length, a size (16:9, 9:16 or 1:1) and a voice if you want them. Prefer to check the plan before anything renders? Turn on Let me review the plan first. Prefer to work from your AI assistant? The Ask your AI assistant tab connects it in one screen, then gives you a ready-to-paste prompt built from the same brief.
Review plan comes next when you asked to review first, or when something needs your attention. Every scene is numbered, with its title, script, type and length, and you can edit, reorder or delete any of them (Undo brings a deleted scene back). Voice, size, music and your brand kit sit in one row above the scenes. Your changes save as you go. If something would stop the video from rendering, a things to check list says what to fix, and Show me jumps to that scene. Next to Render video you see how long the video will be and how many AI minutes it uses. Planning is free. Re-rendering an unchanged video is free. Failed renders don't count against your minutes.
Use it when you know what you want to say but have nothing recorded yet — or to get a skeleton you then attach real footage to.
Import a deck
Turn slides into scenes, one scene per slide, with the speaker notes as the starting script.
Use it for content that already exists as a presentation.
Start from a template
A saved structure — scene layout, voice, brand settings — that you apply to new projects so a series stays consistent. See templates.
Use it from the second video in a series onward.
Sources and the asset library
Whatever you bring in becomes a source, stored in your asset library and reusable across projects. Uploading the same recording twice is never necessary.