LOE Lyric Video Studio
The primary experience now runs inside ChatGPT or Codex as a chatbot-native MCP app. This page stays live as public docs and a fallback planner for the same locked Doubted + M3ntally-iLL treatment, section crossfades, face-safe framing, and local render workflow.
Chatbot UI first
Connect the chatbot-native app at
https://illcoappiverse.vercel.app/api/loe-lyric-video-studio/mcp
in ChatGPT developer mode or Codex so the workflow stays inside the conversation UI.
What is live here
This page gives you a production-facing planner, preset management in-browser, and copy-ready commands for the existing iLLCo-Ai Python renderers.
What stays local
Audio, portrait PNGs, lyric timing files, ffmpeg, and the final MP4 never need a public backend to do the actual job.
Quality rule
Karaoke timing is only final after manual word-level QC. This planner keeps that rule visible so the public page never pretends ASR alone is exact sync.
Build the final lyric-video command
Render checks
Build a portrait proof clip first
Boundary notes
- Use proof clips to confirm framing before spending time on the full export.
- M3ntally-iLL should stay top-biased so the face remains visible in 16:9.
- Do not mark lyric timing exact until a manual spot-check clears it.
Current render artifacts
No fake backend
This is a real Vercel-hosted Appiverse surface, but it does not pretend to be a cloud renderer. It plans, documents, and generates commands for the local iLLCo-Ai workflow.
Current default posture
Keep media private, keep renders local, proof each artist section before the full export, and use manual lyric timing QC before calling the job final.