Utuby

Free, open-source YouTube transcript extractor -- paste a link, get the transcript.

Live at utuby.vercel.app

---

Features

• Instant transcript extraction from any YouTube video URL
• Copy for LLM -- clean flowing text with no timestamps, ready to paste into AI context windows
• Token count estimates so you know if a transcript fits your model's context
• Multiple export formats -- Clean text, Timestamped, SRT, and TXT download
• Transcript search -- find specific moments within a transcript
• Lines / Paragraph toggle -- switch between display modes
• Multi-language captions -- extract transcripts in any available language

---

MCP Integration

Utuby includes a Model Context Protocol (MCP) server that lets LLMs extract YouTube transcripts programmatically. Works with Claude Code, Claude Desktop, and any MCP-compatible client.

Tools

Tool	Description
get_transcript	Full transcript in clean, timestamped, SRT, or summary format
get_video_info	Video metadata, chapters, duration, available captions
search_transcript	Search within a transcript -- returns matching segments with context
get_transcripts	Batch extract from multiple videos at once (up to 10)
get_playlist	Extract all transcripts from a YouTube playlist (up to 25 videos)
get_comments	Fetch top comments for sentiment analysis

Setup

# 1. Clone and install
git clone https://github.com/BEIRUX/Utuby.git && cd Utuby && bun i

# 2. Register the MCP server
claude mcp add utuby -- node ./mcp/server.mjs

That's it. Your LLM can now call any of the tools above directly.

---

Quick Start

# Clone the repo
git clone https://github.com/BEIRUX/Utuby.git
cd Utuby

# Install dependencies
bun i

# Run the dev server
bun dev

Open http://localhost:3000 to use the app locally.

---

Tech Stack

• Next.js 16 (App Router, Turbopack)
• React 19
• Tailwind CSS 4
• Framer Motion
• Radix UI
• @modelcontextprotocol/sdk
• Bun

---

Built by BEIRUX