The YouTube Translate MCP is a specialized tool that allows AI assistants to "read" and understand YouTube videos by accessing their dialogue. It acts as a bridge between video content and large language models, making it possible to extract information from a video without needing to watch it. This makes it an essential utility for anyone who wants to quickly grasp the main points of a tutorial, lecture, or presentation through a simple text-based interface. Beyond basic transcription, this server offers a robust suite of features for deeper content analysis. It can automatically translate video scripts into various languages, generate professional subtitle files in SRT or VTT formats, and produce concise summaries of lengthy recordings. It also features a targeted search capability, enabling users to pinpoint specific mentions of keywords or topics within a video's timeline, which significantly speeds up the research process. For developers and engineers, this tool provides a streamlined way to integrate video data into AI-driven workflows using the Model Context Protocol. Built on Python 3.12, the server supports both stdio and SSE transports, allowing for flexible deployment options through Docker, Smithery, or manual installation. By simply configuring an API key, developers can empower their LLM applications to analyze, translate, and cross-reference YouTube content at scale, effectively turning the world’s largest video platform into a searchable, structured knowledge base.
How to install and configure YouTube Translate MCP
1. Installation Via Smithery (Recommended for Claude Desktop) This command automatically installs the server: bash npx -y @smithery/cli install @brianshin22/youtube-translate-mcp --client claudeManual Installation Requires Python 3.12 or higher. Using uv (recommended): bash uv pip install youtube-translate-mcp Using pip: bash pip install youtube-translate-mcpDocker Installationbash docker build -t youtube-translate-mcp . ---
2. Configuration To use this server, you must have a YOUTUBE_TRANSLATE_API_KEY.
Claude Desktop Configuration Add the following to your claude_desktop_config.json file (typically located at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS): Method 1: Local Development (Using uv)json { "mcpServers": { "youtube-translate": { "command": "uv", "args": [ "--directory", "/ABSOLUTE/PATH/TO/youtube-translate-mcp", "run", "-m", "youtube_translate_mcp" ], "env": { "YOUTUBE_TRANSLATE_API_KEY": "YOUR_API_KEY" } } } }Method 2: Docker-basedjson { "mcpServers": { "youtube-translate": { "command": "docker", "args": [ "run", "-i", "--rm", "-e", "YOUTUBE_TRANSLATE_API_KEY", "youtube-translate-mcp" ], "env": { "YOUTUBE_TRANSLATE_API_KEY": "YOUR_API_KEY" } } } } ---
3. Available Tools The server provides tools to interact with the YouTube Translate API: * Get Transcripts: Retrieve the transcript text of a YouTube video. * Translate Transcripts: Translate existing transcripts into different target languages. * Generate Subtitles: Output video subtitles in SRT or VTT formats. * Create Summaries: Generate concise summaries of the video content. * Search Content: Search for specific keywords or segments within a video. ---
4. Example Prompts * "Get the transcript for the YouTube video at [URL]." * "Summarize the main points of this video: [URL]." * "Translate the transcript of [URL] into French." * "Search for mentions of 'artificial intelligence' in the video [URL]." * "Generate an SRT subtitle file for the video [URL]."
What you can do with YouTube Translate MCP
Use Case 1: Cross-Language Market Research Problem: Researchers often need to analyze product reviews, tech launches, or market trends from foreign markets (e.g., Japan, Korea, or Germany), but manual translation of long videos is time-consuming and often inaccurate with standard auto-captions. Solution: This MCP allows researchers to pull full transcripts of foreign-language videos and translate them into their native language directly within their AI workflow. The AI can then analyze the sentiment and key takeaways of the international audience. Example: A product manager asks Claude: "Get the transcript for this Japanese review of our competitor's new phone, translate it to English, and list the top three features the reviewer liked and the two things they complained about."
Use Case 2: Efficient Content Repurposing for Creators Problem: Content creators and bloggers often want to turn their own (or others') long-form video content into written articles, newsletters, or social media threads, but transcribing the video manually is a bottleneck. Solution: Using the summary and transcript features, the MCP can instantly provide the structured text of a video. An AI can then use this text to draft a blog post or a series of tweets that capture the essence of the video. Example: "Fetch the transcript for my latest 20-minute tutorial video and use it to draft a 1,000-word technical blog post with headings and bullet points for the key steps."
Use Case 3: Deep Technical Search within Long-Form Tutorials Problem: Developers and students often watch 2-3 hour long coding bootcamps or lectures to find a specific 5-minute explanation of a single concept (e.g., "how to configure Webpack"), which usually requires tedious scrubbing through the timeline. Solution: The MCP’s search feature allows users to query specific content within the video text. Instead of watching the whole video, the user can jump straight to the relevant information. Example: "Search through this 4-hour 'Advanced React' video for the specific section where the instructor explains 'Error Boundaries' and provide the transcript for just that part."
Use Case 4: Rapid Accessibility Compliance (SRT
VTT Generation) Problem: Small-scale developers or educational creators need to provide accessible content with subtitles in multiple languages but may not have the budget for professional captioning services. Solution: The MCP can generate industry-standard subtitle files (SRT or VTT) by translating the original transcript into various languages, making it easy to upload them back to YouTube or a private video hosting…