Brainery

Claude Code skill for importing web content, YouTube videos, and PDFs into local RAG system.

🇮🇹 Italiano | 🇨🇳 中文

Overview

Brainery enables Claude Code to import and query web content using AnythingLLM as a local RAG (Retrieval-Augmented Generation) database. Import articles, YouTube transcripts, and PDFs, then query them using natural language.

Key Features:

• 🌐 Web scraping with clean markdown extraction (Crawl4AI)
• 📺 YouTube transcripts with automatic fallback to audio transcription (yt-dlp + Whisper)
• 📄 PDF import with text extraction
• 📁 Local document parsing (PDF, DOCX, TXT, logs) via unstructured-mcp-server
• 🧠 Local RAG database for private, offline content querying (AnythingLLM)
• 💰 Free LLM provider compatible with AnythingLLM (iFlow Platform - 200K context tokens free tier)
• 🐳 Docker-based infrastructure with pre-built images
• 🌍 Multilingual documentation (EN/IT/ZH)

Quick Start

1. Install Docker Containers

git clone https://github.com/Tapiocapioca/brainery-containers.git
cd brainery-containers
docker-compose up -d

Container Stack:

• crawl4ai (port 9100) - Web scraping
• yt-dlp-server (port 9101) - YouTube transcripts
• whisper-server (port 9102) - Audio transcription
• anythingllm (port 9103) - RAG database

2. Configure AnythingLLM

Open http://localhost:9103 and:

1. Create admin account
2. Configure LLM provider (recommended: iFlow Platform - free tier with 200K context tokens)
   • Provider: OpenAI Compatible
   • Base URL: https://vibex.iflow.cn/v1
   • Model: glm-4.6
   • Get API key from: https://iflow.cn
3. Create workspace (e.g., "brainery")
4. Generate API key in Settings → API Keys

💡 Why iFlow? Free tier with excellent multilingual support (EN/IT/ZH), 200K context window, and fast response times. See Installation Guide for detailed setup.

3. Install MCP Servers

Install required MCP servers for Claude Code integration:

# Install AnythingLLM MCP server (use our fork)
npx -y @smithery/cli install @tapiocapioca/anythingllm-mcp-server --client claude

# Install other MCP servers
npx -y @smithery/cli install crawl4ai --client claude
npx -y @smithery/cli install yt-dlp --client claude
\n# Install local document parser
pip install unstructured-mcp-server

📝 Note: We use a custom fork of AnythingLLM MCP server: https://github.com/Tapiocapioca/anythingllm-mcp-server

4. Install Skill

cd ~/.claude/skills
git clone https://github.com/Tapiocapioca/brainery.git

Restart Claude Code to load the skill.

5. Test Import

In Claude Code:

Import this article into Brainery: https://example.com/article

Then query:

What are the main points in the article I just imported?

Usage Examples

Import Web Page

Import this technical guide: https://example.com/docker-guide

Import YouTube Video

Import the transcript from: https://www.youtube.com/watch?v=VIDEO_ID

Batch Import

Import these articles:
1. https://example.com/article1
2. https://example.com/article2
3. https://example.com/article3

Then tell me what the common themes are.

Query Imported Content

What are the key concepts discussed in all the articles I imported today?

Documentation

• Installation Guide - Complete setup instructions
• Usage Examples - Practical workflows and common scenarios
• BRAINERY_CONTEXT.md - Technical implementation details

Other Languages

• Italiano: Installazione | Esempi
• 中文: 安装 | 使用示例

Architecture

Brainery uses a modular architecture with 5 Docker containers:

┌─────────────┐
│ Claude Code │
└──────┬──────┘
       │ MCP Tools
       │
       ├──────────────┬──────────────┬──────────────┐
       ▼              ▼              ▼              ▼
  ┌─────────┐   ┌──────────┐   ┌─────────┐   ┌──────────────┐
  │Crawl4AI │   │ yt-dlp   │   │ Whisper │   │ AnythingLLM  │
  │  :9100  │   │  :9101   │   │  :9102  │   │    :9103     │
  └─────────┘   └──────────┘   └─────────┘   └──────────────┘
   Web scraping  YouTube        Audio          RAG Database
                 transcripts    transcription

System Requirements

• Docker Desktop 20.10+
• Docker Compose 2.0+
• 12GB RAM minimum (16GB recommended)
• ~20GB disk space for containers and models

Port Configuration

Default ports (9100-9104) work out-of-box. To customize, create .env file:

cd brainery-containers
cp .env.example .env
# Edit ports in .env
docker-compose up -d

Troubleshooting

Containers Not Running

docker ps --filter "name=brainery-"
docker-compose restart <service-name>

AnythingLLM "Unauthorized"

Verify API key in .env file and reinitialize MCP client.

Import Fails

1. Check container health: curl http://localhost:9100/health
2. Verify network connectivity
3. Check logs: docker-compose logs <service-name>

See Installation Guide for detailed troubleshooting.

Repository Structure

• brainery (this repo) - Claude Code skill with multilingual docs
• brainery-containers - Docker infrastructure with pre-built images on Docker Hub

Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create feature branch
3. Test thoroughly
4. Submit pull request

License

MIT License - see LICENSE file.

Support

• Issues: GitHub Issues
• Containers: brainery-containers issues
• Documentation: See docs/ for detailed guides

Version

Current version: 1.0.0

See CHANGELOG.md for version history.