Skip to content
SkillMixed787 tokens in use

Video Caption Generator

Transcribe new clips from a Google Drive folder with Whisper and get a social caption and a short YouTube or Facebook title for each.

3,862 stars on the repository MIT Licenseericosiu/ai-marketing-skills · video-caption-generator
npx skills add ericosiu/ai-marketing-skills -s video-caption-generator
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Video Caption Generator costs

23tokens idle
787tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Mixed

Script-ledMixedModel-led

Scripts do part of the work and the model's judgement does the rest.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/process_videos.py.

Touches

shelldetectednetworkdetectedwrites filesdetectedcredentialsdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter8/15
  • Structure19/25
  • Examples12/15
  • Output7/15
  • Size15/15
  • Resources5/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Good
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 34
  • Reaches the network−10

    does this on the machine, and the frontmatter does not say so

    scripts/process_videos.py, line 86
  • Writes files−8

    does this on the machine, and the frontmatter does not say so

    scripts/process_videos.py, line 136
  • Reads credentials−15

    does this on the machine, and the frontmatter does not say so

    scripts/process_videos.py, line 151

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity88
  • Maintenance44
  • Community56
  • Maturity28
  • Momentum40

Video Caption Generator processes the MP4 clips dropped into a Google Drive folder. A Python script downloads each new video, transcribes it with a local Whisper install, and asks Claude for a short first-person caption and a title under 60 characters.

It remembers processed Drive file IDs so a clip is handled once, and it tags clips with identical transcripts as A/B title variants and processes all of them. Results print to the terminal in a Slack-friendly format.

What You Get

  • A transcript for every new clip in the Drive folder
  • A two to four sentence social caption per clip in first person
  • A YouTube and Facebook title under 60 characters per clip
  • A processed_ids.json log so later runs skip clips already handled
  • A tag on clips with matching transcripts that marks them as A/B title variants

Common Use Cases

  • Preparing captions for a batch of short clips before scheduling them
  • Giving an editor transcripts and titles without opening each video
  • Handling A/B title tests where the same audio carries different on-screen titles
  • Running a recurring pass over a folder that fills up during the week

What Makes It Unique

The skill is narrow and ready to run: folder in, copy out. Clips with identical transcripts are treated as A/B variants and all are processed, which suits teams that test on-screen titles. The Drive access goes through a command you configure, so it works with a Drive CLI or wrapper of your choice.

Under The Hood

The agent runs one Python script from the shell. The script calls a Google Drive command line tool that you configure, lists and downloads videos into a temporary folder, runs the local Whisper binary on each file, and posts the transcript to the Anthropic API with your key to get the caption and title. It writes a JSON log of processed IDs in the working directory and prints results to the terminal. The SKILL.md starts with a second, duplicated frontmatter block that describes the same skill with trigger phrases.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers