Can AI replace Zight?

Build a personal replacement for Zight in an empty repository.
Use Tauri 2, React, TypeScript, ffmpeg, and SQLite; do not offer alternative stacks.
The core loop is: capture a screen recording, GIF, or screenshot with optional microphone audio, trim and annotate the result, and export a local file or create a shareable hosted link.
Make the first run work locally with one documented command.
Store all user data locally by default and make export straightforward.
Put secrets in .env, ship .env.example, and never commit credentials.
Offer screen, window, or region capture plus optional microphone audio.
Show an always-visible recording timer and make stop available from the tray.
Create a non-destructive trim view with crop, zoom, cursor highlight, and text callouts.
Render locally through ffmpeg with sensible H.264 and WebM presets.
Generate a poster frame, transcript, and Markdown summary beside each video.
Provide copy-file-path and reveal-in-folder actions instead of mandatory cloud upload.
Include clear empty, loading, success, and recoverable error states.
Add input validation, safe filenames, and graceful handling of unavailable APIs.
Write focused tests for the core transformation and one end-to-end happy path.
Create a README with setup, architecture, permissions, data location, and backup steps.
Do not add accounts, billing, telemetry, analytics, or a hosted control plane.
Deliberately leave out public cloud hosting.
Deliberately leave out viewer identity and engagement analytics.
Deliberately leave out enterprise workspaces and SSO.
Finish by running the tests and listing the exact commands used.