A hybrid Node.js/Python system for the HeadyConnection ecosystem with patented Sacred Geometry architecture, featuring a web-based Admin IDE with AI assistance, real-time build/audit monitoring, and optional remote GPU support.
| Component | Technology | Purpose |
|---|---|---|
heady-manager.js |
Node.js/Express | MCP server, Admin API, static file serving |
src/process_data.py |
Python | Hugging Face inference worker |
src/consolidated_builder.py |
Python | Build orchestration with multi-agent coordination |
admin_console.py |
Python | System audit and health checks |
public/index.html |
React (CDN) | Sacred Geometry UI dashboard |
public/admin.html |
React + Monaco | Admin IDE (file tree, editor, logs, AI panel) |
render.yaml |
Render Blueprint | Infrastructure-as-code deployment |
-
Install dependencies:
npm install pip install -r requirements.txt
-
Set environment variables:
export HEADY_API_KEY="your-api-key" export HF_TOKEN="your-hf-token" export DATABASE_URL="postgresql://..." # Optional
-
Start the server:
npm start # or node heady-manager.js -
Access interfaces:
- Main UI: http://localhost:3300
- Admin IDE: http://localhost:3300/admin
Deploy via render.yaml which uses heady-shared-secrets env group for secrets management.
PORT– Server port (default: 3300)NODE_ENV– Environment (development/production)HEADY_API_KEY– Required for Admin API and HF endpointsHF_TOKEN– Hugging Face API tokenDATABASE_URL– Postgres connection string (optional)HEADY_CORS_ORIGINS– Comma-separated allowed origins
HEADY_ADMIN_ROOT– Repository root for file access (default: repo root)HEADY_ADMIN_ALLOWED_PATHS– Comma-separated allowlist for additional rootsHEADY_ADMIN_MAX_BYTES– Max file size for editing (default: 512 KB)HEADY_ADMIN_BUILD_SCRIPT– Path to build script (default:src/consolidated_builder.py)HEADY_ADMIN_AUDIT_SCRIPT– Path to audit script (default:admin_console.py)
HEADY_ADMIN_ENABLE_GPU– Enable GPU features (true/false)REMOTE_GPU_HOST– Remote GPU server hostREMOTE_GPU_PORT– Remote GPU server portGPU_MEMORY_LIMIT– GPU memory limit in MBENABLE_GPUDIRECT– Enable GPUDirect RDMA (true/false)
The Admin IDE (/admin) provides a complete web-based development environment:
- File Browser – Navigate project files with allowlisted roots and safe path resolution
- Monaco Editor – Full-featured code editor with syntax highlighting (Python, JSON, YAML, JS)
- Multi-Tab Editing – Work on multiple files simultaneously with Ctrl+S save
- Real-Time Logs – Build/audit logs streamed via Server-Sent Events
- Settings Panel – Configure GPU and system settings (local development only)
- AI Assistant – Integrated AI code assistance (MCP/Copilot integration)
GET /api/admin/roots– List allowed file system rootsGET /api/admin/files?root=&path=– Browse directory contentsGET /api/admin/file?root=&path=– Read file contentPOST /api/admin/file– Write file (with SHA-256 conflict detection)POST /api/admin/build– Trigger build scriptPOST /api/admin/audit– Trigger audit scriptGET /api/admin/ops– List operationsGET /api/admin/ops/:id/status– Get operation statusGET /api/admin/ops/:id/stream– SSE log stream for operationGET /api/admin/config/render-yaml– Get render.yaml configurationGET /api/admin/config/mcp– Get MCP configuration (secrets masked)GET /api/admin/settings/gpu– Get GPU settings (secrets masked)POST /api/admin/assistant– AI assistant for code editingPOST /api/admin/lint– Code lintingPOST /api/admin/test– Run tests
POST /api/hf/infer– Generic HF inferencePOST /api/hf/generate– Text generationPOST /api/hf/embed– Text embeddings
GET /api/pulse– System status and Docker infoGET /api/health– Simple health check
python src/consolidated_builder.py --project-root /path/to/project --output build-info.json# Full audit
python admin_console.py --project-root /path/to/project
# Specific checks
python admin_console.py --check health
python admin_console.py --check structure
python admin_console.py --check deps
python admin_console.py --check security# Python tests
pytest
# Node.js tests (when implemented)
npm testTo run this project on Google Colab and utilize free GPU resources:
- See COLAB_PROTOCOL.md for detailed instructions
- Use the
notebooks/Heady_Colab_Protocol.ipynbnotebook
This project is optimized for GitHub Copilot with MCP integration:
.github/copilot-instructions.md– Project overview and Quiz Protocol for documentation.github/copilot-mcp-config.json– MCP server definitions.github/workflows/copilot-setup-steps.yml– Setup workflowmcp_config.json– Local development MCP servers
- filesystem – File system operations
- sequential-thinking – Reasoning chains
- memory – Persistent memory storage
- fetch – Web requests
- postgres – Database operations
- git – Git operations
- puppeteer – Web automation
- cloudflare – Cloudflare API
All documentation follows the Quiz & Flashcard Methodology (see .github/copilot-instructions.md).
- All admin endpoints require
HEADY_API_KEYauthentication - Secrets managed via environment variables, never hardcoded
- Rate limiting enabled on API endpoints (120 req/min default)
- CORS configured for allowed origins
- File operations restricted to allowlisted roots
- SHA-256 conflict detection for file writes
- Fork the repository
- Create a feature branch
- Follow the code style guidelines in
.github/copilot-instructions.md - Ensure all tests pass
- Submit a pull request
See LICENSE file.