Skip to content
Kitbay
Skill

ai-presenter-video

MITUpdated on Nov 11, 2026262,606GitHub starsIncludes scripts

An optional skill for Hermes Agent that turns a script and a photo of a presenter into a talking-presenter video.

What the skill teaches

An optional skill for Hermes Agent that turns a script and a photo of a presenter into a talking-presenter video. The agent sets up a job folder, records the user's confirmation of image rights, adult status, remote upload and any voice cloning, and runs a preflight check. It then narrates the script with text-to-speech, checks the audio with speech recognition, runs a short paid pilot of the lip-synced presenter (for example FAL image-to-video models) before the full run, edits with ffmpeg, adds captions, and finalises loudness, encodes and a contact sheet for review. Paid calls need explicit approval, and it stops after three rejected attempts. It needs ffmpeg and Python 3 on Linux or macOS.

Capabilities

  • Narrate a script with text-to-speech and check timing
  • Generate a lip-synced presenter video from one image
  • Add captions and edit with ffmpeg
  • Normalise loudness and export share-ready files