Say It Ahead journal
Notes from
building the practice.
Straightforward reports about live voice, conversation design, cost, privacy, and what the product still gets wrong.
Why the AI character would not calm down, and how I fixed it
A plain-language look at the four-step conversation model behind Say It Ahead, why early role-plays stayed angry, and how live progress differs from feedback.
8 minute read→02 / Build notesWhat I learned building a real-time AI conversation practice tool
A practical look at live voice latency, character behavior, feedback, cost controls, and the parts of AI role-play that still need work.
8 minute read→