Use the free Local Voice Add-on included with my Make Video Skill purchase to make my first video with my chosen voice. Chosen American voice: Heart (af_heart). Use the snappy 1.25x pace. Do the routine free setup and checks, then give me links to the wide and vertical MP4s. Use simple language.

Locate my extracted local-video-kit and Local-Voice-Add-on-1.3.0 folders. Keep both complete. Read their SKILL.md, the core SETUP-WITH-AI.md, and the add-on INSTALL.md, KOKORO.md and production-bridge instructions. Use the actual included scripts, never a similarly named package. Work in a fresh named output folder; preserve existing videos and voices.

Check this computer and the documented platform scope. The complete production route was tested on Apple Silicon macOS with 32 GB memory; that is not a verified minimum or a Windows/Linux/Intel Mac guarantee. Check Node.js 20.9+ and npm, Python 3.11, 3.12 or 3.13 (3.12 tested), FFmpeg and FFprobe with libx264 and the ass caption filter. Install only missing free prerequisites from reviewed official sources or an existing trusted package manager where normal permissions allow. Do not buy anything, accept legal/security prompts for me, silently use administrator access, change assistant settings or disable security. Tell me any exact action that needs my own involvement.

I authorize the toolkit's exact locked dependencies and the add-on's pinned Kokoro environment, model and preset downloads. Inspect their files first. Use node scripts/assistant-setup.cjs check, install the core dependencies if needed with node scripts/assistant-setup.cjs install, and check again. Install only the selected Kokoro engine for this film. Do not download Qwen or unrelated models. Keep downloaded models and private machine configuration out of files intended for sharing.

Run python3 "ADDON/voice_video.py" doctor --engine kokoro --voice af_heart --kit "CORE", replacing ADDON and CORE with the actual extracted folder paths. Resolve reported prerequisites before rendering. Then run python3 "ADDON/voice_video.py" make --engine kokoro --voice af_heart --kit "CORE" --pace 1.25 --out "NEW-UNIQUE-OUTPUT-FOLDER" --install. Use the default full-HD output. On Windows use the supported Python launcher and explain any unverified complete-workflow step before proceeding.

This first test uses the included original nine-scene example. It generates fresh scene narration with my chosen American English voice and uses the core's native renderer for wide and vertical MP4s. Its captions are estimated word and phrase cues based on final paced scene audio, not forced alignment. Do not reuse an old starter's unrelated speech or captions, and do not claim native phoneme-accurate lip sync.

Check the two actual MP4s, their duration, full decode, captions, source and bridge verification records. Open a representative frame from each. Show the clickable local movie links and exact editable project folder. Tell me to listen to the whole result and review caption timing; machine checks are not human approval.

For my own topic later, help me author a new native story JSON from the included first-film.story.json, with a narration string per scene, then run the same bridge with --story and a new output folder. Do not claim that this bridge automatically generates every style from arbitrary text. Do not publish, upload private files or use a paid service.