whatisgithub

What is watch-skill?

oxbshw/watch-skill — explained in plain English

Analysis updated 2026-05-18

224PythonAudience · developerComplexity · 3/5LicenseSetup · moderate

In one sentence

A local-first tool that gives AI agents the ability to watch, remember, and verify video, including their own screen recordings.

Mindmap

mindmap
  root((Watch Skill))
    What it does
      Watches videos and streams
      Builds searchable index
      Verifies agent's own work
    Tech stack
      Python
      MCP tools
      Whisper
      Vision AI providers
    Use cases
      Search video content
      Cite timestamped evidence
      Check agent browser tasks
    Audience
      AI agent developers
      Claude Code users

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Let an AI coding agent record and review its own browser session to verify a task was completed correctly.

USE CASE 2

Search across a personal library of watched videos and meetings for a specific moment with a timestamp.

USE CASE 3

Summarize a long video or livestream and ask follow-up questions about specific parts of it.

What is it built with?

PythonMCPClaude CodeWhisperOllama

How does it compare?

oxbshw/watch-skillanybackup-ai/anybackupdjango-haystack/celery-haystack
Stars224222227
LanguagePythonPythonPython
Last pushed2026-07-06
MaintenanceActive
Setup difficultymoderatehardmoderate
Complexity3/54/53/5
Audiencedeveloperops devopsdeveloper

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 30min

Local transcription and search need no API key, but visual question answering requires connecting a paid vision AI provider or a local Ollama model.

MIT license: free to use, copy, modify, and share, including for commercial purposes, as long as the original copyright notice stays included.

So what is it?

Watch Skill is a tool that gives AI coding agents the ability to watch video, understand what happened in it, and check their own work by recording and reviewing it. It turns videos, live streams, meeting recordings, and screen recordings of an agent's own actions into a searchable, timestamped record that an agent can search, ask questions about, and cite specific moments from. The project describes three main abilities it adds to an agent. Watch lets it process video from more than 1,800 websites, live streams, local video files, and recordings of browser or desktop sessions, extracting frames, on screen text, and transcripts. Remember builds a persistent, searchable index of everything analyzed, so an agent can ask a new question about a video later without re-downloading or re-processing it, and can find related material across every video it has seen before. Verify, referred to in the README as THE LOOP, is meant for when the video is a recording of the agent's own work: it records the session, checks the result against plain language criteria the user sets, guides a fix if something is wrong, and produces before and after evidence that the fix worked. It is installed either as a plugin inside Claude Code, or through an install script on macOS, Linux, or Windows, and is also usable through a command line tool, a REST API, and 23 tools exposed over the MCP protocol so many different AI coding agents and editors can use it. Transcription and search can run entirely on a local machine without any API key, while answering visual questions about a video requires connecting a vision capable AI provider such as Anthropic, Gemini, OpenAI, OpenRouter, or a locally run Ollama model. It is written mostly in Python, released under the MIT license, and the project includes its own published cost and accuracy benchmarks rather than relying on unverified marketing claims.

Copy-paste prompts

Prompt 1
Walk me through installing Watch Skill as a Claude Code plugin and running the setup command.
Prompt 2
Explain how THE LOOP records, critiques, fixes, and verifies an agent's browser flow according to this README.
Prompt 3
Help me use the watch-skill CLI to summarize a YouTube video and then ask a follow-up question about it.
Prompt 4
What vision providers can I connect for visual question answering, and how do I keep everything running locally instead?

Frequently asked questions

What is watch-skill?

A local-first tool that gives AI agents the ability to watch, remember, and verify video, including their own screen recordings.

What language is watch-skill written in?

Mainly Python. The stack also includes Python, MCP, Claude Code.

What license does watch-skill use?

MIT license: free to use, copy, modify, and share, including for commercial purposes, as long as the original copyright notice stays included.

How hard is watch-skill to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is watch-skill for?

Mainly developer.

Open on GitHub → Ask about another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.