Manual

Getting started

Installation, permissions, your first script and the first run with voice tracking.

Installation

TalkPatch is a native app for macOS 14 or newer. Drag TalkPatch.app into your Applications folder and open it. The first time it launches, macOS asks for three permissions:

Permission What for
Camera shows your camera image behind the prompter text and records video
Microphone level meter, voice tracking and audio recording
Speech recognition only for the Apple model; the local models don’t need it

All of it runs on your Mac. TalkPatch sends no audio, no video and no text to any server.

The window

  • Stage (center): prompter text on the camera image or on a background color. The blue line is the reading line. The chip below it shows chapter, time and progress, and you can drag the progress bar.
  • Scripts (left, ⇧⌘L): library and chapters of the current script.
  • Inspector (right, ⌥⌘I): text, speech, notch. Font, color, text width, margins, mirroring, ring light, chapter headings, speech model.
  • Studio bar (bottom): camera and microphone chips, transport, voice tracking, recording. Position (bottom, top, left, right) under Settings → General.

The scripts panel and the Inspector float above the stage. The text stays where it is when you show or hide them.

Your first script

  1. New text (⌘N) opens the editor, or pull in a .txt/.md/.rtf file with Import text … (⌘O).
  2. ## Chapter name starts a chapter, > Stage direction stays in the editor and never appears in the prompter text. Everything else is prompter text.
  3. Stop editing (⌘E) shows the prompter text.

The library saves itself after every change.

The first run

  • Space starts and pauses. The countdown before the start can be switched off in the Speech view.
  • S stops, Home jumps to the beginning.
  • ←/→ change the chapter, ↑/↓ go one line back or forward.
  • Switch on voice tracking in the studio bar. TalkPatch downloads the selected model the first time (Nemotron multilingual by default). After that the reading line follows your voice, and when you pause, the text waits.

Every shortcut can be reassigned under Settings → Keys, each with a tap and a hold.