How voice learning works

Sparky builds a structured voice profile from your transcripts: tone, sentence style, vocabulary, hooks, banned words, and how you open posts. Onboarding role and goals still shape the angle of each post.

When your profile updates

• After every project transcript is stored (including transcript file uploads).
• For thought leader and expert guest projects, only your selected speaker lines are analyzed.
• When you approve a project post you edited - Sparky may add style correction rules from the diff.

Voice Confidence score

The top bar show Voice Confidence as a percentage. It answers one question: how much transcript material has Sparky already used to learn how you write and speak?

The score is not a grade on post quality. It is a depth meter. More processed transcripts mean a richer profile and a higher number, up to a built-in cap of 95%.

How the score is calculated (C(n))

Behind the scenes SparkVox uses a simple curve called C(n). The n is how many project transcripts have been counted toward your voice profile (each new project transcript counts once; re-processing the same project does not).

Think of it like filling a glass: the first transcript moves the score a lot. The second moves it again, but a bit less. By the time you have several episodes in the system, each new one still helps, but the number inches forward more slowly. That is intentional - early projects teach Sparky the basics fast; later ones refine what is already there.

The curve never quite reaches 100%. It levels off near 95% so the score always leaves room for more learning as you keep publishing. Deleting a project does not lower the score; transcript counts only go up.

How profiles merge

Each new transcript strengthens patterns Sparky already knows, adds new ones, and drops anything clearly contradicted. Your profile is never replaced unless you fire Sparky.