video-research-mcp

Overview

video-research-mcp is a comprehensive MCP server that extends Claude Code's capabilities with video processing, deep research, and media production tools. It includes 34 tools for video analysis and research, 15 for explainer video production, and 2 for scene generation. The server leverages Gemini 3.1 Pro for AI capabilities and ElevenLabs for text-to-speech features. It supports local file processing, YouTube analysis, web search, and academic paper research with evidence tiering. The system includes a knowledge store using Weaviate for persistent, cross-project information retrieval.

Try asking AI

After installing, here are 6 things you can ask your AI assistant:

• video_analyze

  Analyze video content and extract insights

  Try:Analyze this meeting recording for action itemsTry:Summarize this tutorial video
• research_deep

  Conduct deep research with evidence-tier labeling

  Try:Research HNSW index parameters for high-dimensional embeddingsTry:Find recent developments in quantum computing
• web_search

  Search the web using Gemini with source citations

  Try:Search for latest MCP protocol developmentsTry:Find information about climate change impacts
• knowledge_search

  Search stored knowledge in Weaviate

  Try:What do I know about Kubernetes?Try:Find my previous research on neural networks
• content_analyze

  Analyze documents, PDFs, or text content

  Try:Analyze this research paper for key findingsTry:Summarize the contents of this directory of documents
• explainer_create

  Create an explainer video project

  Try:Create a project for explaining quantum computingTry:Set up a video explainer for our new product
• explainer_generate

  Generate content for an explainer video

  Try:Generate a script for my quantum computing explainerTry:Create visual content for the project introduction
• research_web

  Launch web research agent with followup capabilities

  Try:Start research on AI ethics and report backTry:Research market trends for electric vehicles
• video_metadata

  Extract metadata from YouTube videos

  Try:Get metadata for this tutorial videoTry:Analyze this YouTube video for key information
• knowledge_ingest

  Import external structured knowledge into Weaviate

  Try:Ingest this research paper into my knowledge baseTry:Add these industry reports to my knowledge store
• research_paper_search

  Search for academic papers and publications

  Try:Find recent papers on transformer architecturesTry:Search for research on climate change modeling
• agent_generate_scenes

  Generate multiple scenes in parallel for video content

  Try:Generate scenes for a documentary about space explorationTry:Create scene concepts for a product demo video

Comparable tools

Installation

npx video-research-mcp@latest
export GEMINI_API_KEY="your-key-here"

{
  "mcpServers": {
    "video-research": {
      "command": "uvx",
      "args": ["video-research-mcp"],
      "env": { "GEMINI_API_KEY": "${GEMINI_API_KEY}" }
    }
  }
}

FAQ

What file formats are supported for analysis?

The server supports MP4, MOV, AVI for video; PDF, TXT, MD, HTML, XML, JSON, CSV for documents; and direct YouTube URLs.

How does the knowledge store work?

When Weaviate is connected, research findings, video analyses, and content are automatically stored and made searchable across projects and sessions without it, data persists per-session only.

Compare video-research-mcp with