A workflow that turns an outline into a full script written for the spoken voice, in your rhythm and your words.
Category
Content Production
Return
~2 hrs / week
Stack
Claude · Content Plan
The Problem
Text that reads well often speaks badly. A script written like an article trips the tongue, buries the breathing points and loses the listener two minutes in. Writing for the spoken voice is its own craft, and doing it from scratch for every video is the slowest step in the whole content pipeline.
The Solution
A scripting workflow tuned for speech. You bring an outline or a raw idea, and the system returns a full script in your rhythm, with natural transitions, breathing points and the repetition that spoken teaching needs. Recording starts the same day the idea does.
How It Works
- An idea or outline enters from the Content Plan, with its pillar and format attached.
- Claude drafts the script from your Personality Definition and Vocabulary Signature, written in speech rhythms rather than page rhythms.
- Spoken-delivery rules apply: flowing sentences, verbal signposts, key points repeated with fresh angles for listeners.
- The draft returns structured for recording, with sections you can deliver in single takes.
- Approved scripts flow onward, to the camera, the Video Avatar or the Voice Clone.
Built With
- The Content Plan as the source of ideas and slots
- The Personality Definition and Vocabulary Signature for voice
- Spoken-delivery rules encoded into the drafting pass
- A recording-ready script template
It sits between the Content Plan and the production layer, feeding scripts to the Video Avatar and the Voice Clone as they come online.
What You Get
- Scripts that sound like you talking, because they were written to be spoken
- The blank-page stage of every video removed
- A consistent teaching rhythm across your whole catalog
The Numbers
One to two days to encode the delivery rules and templates. For anyone producing a video or lesson a week, it returns about two hours per piece of drafting and rework, and the scripts stop sounding read.