JAEGIS AI Web OS transforms architectural documentation into complete, production-ready applications using AI-powered analysis and code generation.
- Node.js 16+ (for CLI wrapper)
- Python 3.8+ (for core processing)
- AI Provider API Key (OpenAI, Anthropic, or local model)
# Interactive mode with guided setup
npx jaegis-ai-web-os interactive
# Direct build from document
npx jaegis-ai-web-os build --base ./architecture.docx --output ./my-app# Install globally via npm
npm install -g jaegis-ai-web-os
# Verify installation
jaegis-ai-web-os --version# Clone repository
git clone https://github.com/usemanusai/JAEGIS-AI-Web-OS.git
cd JAEGIS-AI-Web-OS
# Install dependencies
npm install
pip install -r requirements.txt
# Run locally
node cli.js --help# AI Provider Configuration (choose one or more)
export OPENAI_API_KEY="your-openai-api-key"
export ANTHROPIC_API_KEY="your-anthropic-api-key"
# Optional: Preferred provider
export MCP_PREFERRED_AI_PROVIDER="openai" # or "anthropic", "auto"
# Optional: Processing configuration
export MCP_MAX_CHUNK_SIZE="4000"
export MCP_CACHE_ENABLED="true"
export MCP_LOG_LEVEL="INFO"Create mcp_server.yaml in your project directory:
# AI Provider Settings
ai:
preferred_provider: "openai"
request_timeout: 120
max_retries: 3
# Document Processing
processing:
max_chunk_size: 4000
chunk_overlap_size: 200
# Caching
cache:
enabled: true
max_size_mb: 1024
ttl_hours: 24
# Logging
logging:
level: "INFO"
file_enabled: true# Launch interactive mode
jaegis-ai-web-os interactive
# Follow the guided prompts:
# 1. Enter project name
# 2. Select architecture documents
# 3. Configure AI provider
# 4. Choose framework preference
# 5. Review and confirm
# 6. Watch your project being built!# Basic build from architecture document
jaegis-ai-web-os build \
--base ./docs/architecture.docx \
--output ./my-project
# Enhanced build with specific AI provider
jaegis-ai-web-os build \
--base ./specs/requirements.md \
--output ./my-app \
--ai-provider openai \
--enhanced
# Dry run to preview what will be generated
jaegis-ai-web-os build \
--base ./design.pdf \
--output ./preview \
--dry-run \
--plan-only- Microsoft Word (.docx) - Full structure preservation
- PDF (.pdf) - Text extraction with table detection
- Markdown (.md) - Native support with metadata
- Plain Text (.txt) - Automatic encoding detection
- PowerPoint (.pptx) - Slide content extraction
- Excel (.xlsx) - Data and table extraction
- HTML (.html) - Structure-aware processing
- Size: Up to 100MB per document
- Pages: Up to 500 pages (recommended: 50-100 pages)
- Language: English (primary), other languages supported
- Structure: Clear headings and sections improve results
- Next.js 15: Full-stack React with TypeScript and Tailwind CSS
- React 18: Modern React with Vite and TypeScript
- Vue.js: (Coming soon) Progressive framework
- Angular: (Coming soon) Enterprise applications
- Python: CLI applications with Click and modern tooling
- Django: Web applications with REST API support
- FastAPI: High-performance APIs with documentation
- Node.js/Express: (Coming soon) JavaScript backends
Create a document with:
- Project Overview: What you're building
- Technical Requirements: Technologies, frameworks, features
- Architecture: System design, components, data flow
- Dependencies: External services, databases, APIs
# Interactive mode (recommended for first time)
npx jaegis-ai-web-os interactive- Project Information: Name, description, output directory
- Document Selection: Choose your architecture document(s)
- Configuration: AI provider, framework preference
- Preview: Review the generated build plan
- Generation: Watch your project being created
# Navigate to your project
cd ./output/your-project-name
# For Next.js projects
npm install
npm run dev
# For Python projects
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install -r requirements.txt
python main.py# Input: requirements.docx (web app specifications)
# Output: Complete Next.js application
jaegis-ai-web-os build \
--base ./requirements.docx \
--output ./web-app \
--template-preference nextjs \
--ai-provider openai# Input: api-design.md (API specifications)
# Output: FastAPI service with documentation
jaegis-ai-web-os build \
--base ./api-design.md \
--output ./api-service \
--template-preference fastapi \
--enhanced# Input: cli-specs.pdf (command-line tool requirements)
# Output: Python CLI application
jaegis-ai-web-os build \
--base ./cli-specs.pdf \
--output ./cli-tool \
--template-preference python \
--ai-provider anthropic# Solution: Set API key
export OPENAI_API_KEY="your-api-key"
# or
export ANTHROPIC_API_KEY="your-api-key"# Check document format and size
jaegis-ai-web-os analyze ./your-document.docx
# Try with different format
# Convert to PDF or Markdown if needed# Run in dry-run mode to see issues
jaegis-ai-web-os build --base ./doc.docx --dry-run
# Check logs for details
tail -f logs/mcp_server.log# General help
jaegis-ai-web-os --help
# Command-specific help
jaegis-ai-web-os build --help
jaegis-ai-web-os interactive --help# Check system status and configuration
jaegis-ai-web-os status
# Clear cache if needed
jaegis-ai-web-os clear-cache# Enable verbose logging
jaegis-ai-web-os build --base ./doc.docx --verbose --debug- Configuration Guide - Advanced configuration options
- API Documentation - Python API reference
- Examples - Comprehensive examples and tutorials
- Architecture - System design and components
- Custom Templates - Create your own templates
- AI Provider Configuration - Multi-provider setup
- Performance Optimization - Optimize for large documents
- Enterprise Deployment - Production deployment guide
- GitHub Issues - Bug reports and feature requests
- GitHub Discussions - Community discussions
- Contributing Guide - How to contribute
You're now ready to transform your architectural documentation into production-ready applications with JAEGIS AI Web OS!
Pro Tip: Start with the interactive mode to get familiar with the system, then move to command-line mode for automation and CI/CD integration.