GitHub radar
Codex Skill: AI Presenter Video from a Script
A provider-neutral Codex skill that turns a script and an authorized portrait photo into a finished presenter video — voiceover, digital human, lip-sync, captions, and final cut are all handled by the agent.
Lanshu Create AI Presenter Video is a Codex skill for producing presenter (digital human) videos. You provide a topic or finished script and an authorized portrait photo — the skill writes or refines the copy, synthesizes or clones a voice, generates a digital human video, syncs lip movement to the audio track, adds captions and keyword animations, edits the timeline, and runs a quality check. The workflow fixes the full voiceover first, then uses it as the timing backbone for everything else — reducing lip drift and cut mismatches. Default output is a 9:16 1080×1920 30 fps video, 45–75 seconds. Before any paid step the agent discloses cost and waits for approval. The project is provider-neutral: it selects whichever video, TTS, and lip-sync services are available in the current Codex environment.
Why a vibe-coder should care
The skill collapses a multi-step production workflow — scripting, voice recording, digital human rendering, editing — into a single agent conversation. Explicit cost-confirmation gates mean you won't accidentally spend money or receive a broken clip without warning.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install this Codex skill: https://github.com/cclank/lanshu-create-ai-presenter-video — clone it into ~/.codex/skills/ per the README, then create a 30-second 9:16 video on the topic 'How to use AI for work' with captions and show me the finished file. Ask me if you need API keys for voice or video generation.
Runs on a regular laptop — you'll need Python and FFmpeg. Heavy video and voice generation happens in the cloud; you'll need API keys for the services you choose.
Open on GitHub▌ More finds