Skip to main content
Framesail AI
Docs

Getting started

How Framesail works

Framesail turns a brief or script into a finished long-form video through a pipeline of ordered stations: script → script scan → reference images → voiceover → storyboard → segment renders → scenes → export. Each station produces something you can review, edit, and re-run before the next station consumes it, and the end result is a finished video.

The stations

In the app, these appear as five steps on every project: Script, Characters & Assets, Voiceover, Storyboard, and the Editor.

Steps unlock in order

Each step unlocks when the previous step's output is complete — the script step needs a style, the voiceover step needs every asset to have a reference image, and so on. The same order is enforced server-side, so API and MCP clients get the same protection as the web app. The full unlock ladder is in Make your first video, end to end.

Iterative by design

The pipeline is not a one-way conveyor. Every station stays editable after it runs: revise the script, regenerate a single reference image, re-render one voice block, or rewrite one storyboard segment — without touching the rest. Regenerating individual pieces at any station is normal, expected workflow, not a recovery path.

Generation runs as background jobs. You can kick off a step, leave the page, and come back — each job reports its status and any errors, and the project shows a step-by-step progress report.

What you get at the end

The export is a flat, finished video, rendered in the cloud. There is no project-file or editable-timeline export — drop the MP4 into Premiere, DaVinci, or any editor like any other clip if you want to cut further.

There is no free tier — every plan is paid. Plan names and what each unlocks are on the pricing page, and common questions are answered in the FAQ.