youtube-transcript-mcp: Instructions file for Claude Code

CLAUDE.md

youtube-transcript-mcp CLAUDE.md is an instructions file for Claude Code from yunlinwu/youtube-transcript-mcp. It costs 362 tokens per session, scanned A, original, MIT.

Developer instructions for a Python MCP server that gives an assistant access to YouTube transcripts.

In plain words
What is it for?
Setting up, running, testing, and maintaining the YouTube transcript server and its transcript-search tools.
Why use it?
They explain how the server extracts captions, optionally uses speech recognition when captions are unavailable, handles errors, and is tested.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md.

This is yunlinwu/youtube-transcript-mcp's own configuration. It tells Claude Code how to work on youtube-transcript-mcp itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything youtube-transcript-mcp configures →

Reuse

Borrowing it

Nothing to install: this file belongs to yunlinwu/youtube-transcript-mcp. 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/yunlinwu/youtube-transcript-mcp/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/yunlinwu/youtube-transcript-mcp

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.

agentmods badge for youtube-transcript-mcp CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/yunlinwu/youtube-transcript-mcp/claude-md.svg)](https://agentmods.dev/instructions/yunlinwu/youtube-transcript-mcp/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/yunlinwu/youtube-transcript-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/yunlinwu/youtube-transcript-mcp/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 362 This file is loaded in full into every session.
When invoked 362 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00362 $0.00362
Opus 5 $0.00181 $0.00181
Sonnet 5 $0.00072 $0.00072
Haiku 4.5 $0.00036 $0.00036

Measured 7d ago against content hash c0da4d32c506, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

youtube-transcript-mcp 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 7d 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 · 45 lines

What it actually says

YouTube Transcript MCP Server

What this is

An MCP server (Python) that gives Claude access to YouTube video transcripts. Two tools: get_transcript_tool and search_transcript.

Architecture

server.py    — MCP tool definitions (FastMCP), error handling, response formatting
transcript.py — extraction logic: yt-dlp captions → VTT parsing → Segment list
                Whisper fallback when captions unavailable (optional dep)

Dev setup

python3 -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"          # core + pytest
pip install -e ".[whisper]"      # optional Whisper fallback

Running tests

pytest tests/ -v

Running the server

youtube-transcript-mcp           # stdio mode (used by Claude Desktop)

Key decisions

  • Captions first, Whisper fallback: captions are fast and free; Whisper requires downloading audio + ~140MB model. Whisper is an optional dependency.
  • Tools return {"error": "..."} on failure instead of raising — Claude gets a useful message it can relay to the user.
  • search_transcript merges overlapping context windows so nearby matches don't produce redundant output.
  • Segments include both raw seconds and human-readable timestamps ("1:30:05").

Dependencies

  • mcp[cli] — MCP SDK (FastMCP)
  • yt-dlp — YouTube downloading/metadata/subtitles
  • openai-whisper — speech-to-text fallback (optional)
  • ffmpeg — required by Whisper (system install)
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.

  1. 7d ago First seen · 45 lines · 362 tokens per session scan A c0da4d32c506

Subscribe to this mod's changes

youtube-transcript-mcp CLAUDE.md is an instructions file published in the GitHub repository yunlinwu/youtube-transcript-mcp (0 stars, last pushed 1mo ago), licensed MIT. It adds 362 tokens to every session, about $0.0018 per session on Opus 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.