What is youtube-summarize CLAUDE.md?

Project instructions for youtube-summarize, an MCP server that retrieves YouTube transcripts and can return summarization instructions. MCP is a standard way for an AI client to call tools provided by a server.

Use them to install dependencies, start the server, inspect transcripts, list transcript languages, search videos, or work within the repository's conventions.

They give an agent the project context, commands, architecture, and available tools needed to work on or run the server correctly.

How zlatkoc/youtube-summarize configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to zlatkoc/youtube-summarize. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/zlatkoc/youtube-summarize/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/zlatkoc/youtube-summarize

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/zlatkoc/youtube-summarize/claude-md/github.svg)](https://agentmods.dev/instructions/zlatkoc/youtube-summarize/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/zlatkoc/youtube-summarize/claude-md"><img src="https://agentmods.dev/badge/instructions/zlatkoc/youtube-summarize/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/zlatkoc/youtube-summarize/claude-md"><img src="https://agentmods.dev/badge/instructions/zlatkoc/youtube-summarize/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 922 This file is loaded in full into every session.

When invoked 922 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00922	$0.00922
Opus 5.5	$0.00369	$0.00369
Sonnet 5.5	$0.00184	$0.00184
Haiku 4.5	$0.00092	$0.00092

Measured 2mo ago against content hash 1c21e960745f, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

youtube-summarize CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 83 lines

How it starts

The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

MCP (Model Context Protocol) server that retrieves YouTube video transcripts and optionally summarizes them. It uses the youtube_transcript_api Python library directly and exposes its functionality as MCP tools.

Key Commands

# Install dependencies
uv sync

# Run the MCP server (stdio transport)
uv run mcp run main.py

# Run the MCP dev inspector (web UI for testing)
uv run mcp dev main.py

Architecture

Entry point: main.py — defines the MCPServer instance and all MCP tools
Transport: stdio-based MCP server (launched via mcp run)
Library: youtube_transcript_api — used directly as a Python library (not as a subprocess/CLI)

MCP Tools

get_transcript — fetch a YouTube video's transcript in a specified format (json, pretty, text, webvtt, srt), with optional language selection
list_transcripts — list available transcript languages for a video
summarize_transcript — fetch transcript and return it with summarization instructions for the LLM client to act on
search_videos — search YouTube via yt-dlp (no ads/recommendations/personalization), with sort (relevance/date/views/rating) and upload-date/duration filters encoded as YouTube's sp protobuf parameter
get_video_metadata — fetch full metadata (title, description, channel, upload date, duration, views, chapters, tags, etc.) for a video via yt-dlp
list_playlist_videos — list a playlist's videos (titles, IDs, channels, durations, views) via yt-dlp's flat extraction, with sorting and a limit

Design Decisions

Uses youtube_transcript_api as a Python library directly (structured data, proper exceptions, no subprocess overhead)
Formatter classes from youtube_transcript_api.formatters handle output formatting (JSON, Text, SRT, WebVTT, PrettyPrint)
The summarization tool returns the transcript with a prompt/instruction for the LLM to summarize, rather than calling an external LLM API itself (the MCP client/LLM handles summarization)
Video IDs are extracted from full YouTube URLs when provided (supports youtube.com/watch?v=, youtu.be/, youtube.com/embed/, youtube.com/shorts/, and bare 11-char IDs)
Tools return error strings rather than raising exceptions — the LLM client can act on them

Read the full file on GitHub · 83 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2mo ago First seen · 83 lines · 922 tokens per session scan A 1c21e960745f

Subscribe to this mod's changes

youtube-summarize CLAUDE.md is an instructions file published in the GitHub repository zlatkoc/youtube-summarize (6 stars, last pushed 1mo ago), licensed MIT. It adds 922 tokens to every session, about $0.0037 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

LLM Research anthropic claude claude-desktop mcp mcp-server

youtube-summarize CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
youtube-summarize CLAUDE.mdzlatkoc/youtube-summarize	Instructions file	6	1mo ago	MIT	A	922	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	9,719	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	yesterday	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	yesterday	MIT	A	4,680	Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	5,516	GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	yesterday	MIT	A	3,093	GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness	Instructions file	274k	2d ago	MIT	A	4,313	Codex, OpenCode

youtube-summarize: Instructions file for Claude Code