Ai Automation Prototype

Unlike typical automation tools that simply schedule posts, this app captures the reasoning behind every action and turns it into reusable intelligence. It doesn’t just execute – it remembers, recombines, and evolves.

🧠 Self-Generative Orchestration

• Flow Recombination – New features emerge by combining existing flows (code growth = 0%)
• DSL Execution – Natural language → slot-based intent → deterministic DAG
• Graph RAG – Vector + graph search that stores relationships, not data
• Adaptive Scoring (Adapter Bonus) – Learns how to chain actions automatically

Maestro doesn’t grow by adding new code – it evolves through new combinations.

⚙️ Architecture

• Backend: FastAPI + PostgreSQL (pgvector) + Celery + Redis + SeaweedFS
• Frontend: React 19 + Vite + Zustand + shadcn/ui
• AI Layer: Hugging Face multilingual-e5 embeddings
• Graph RAG: 9 node types × 7 edge types – self-organising knowledge graph
• Adapters: Threads & Instagram (Meta Graph API integration)

🚀 Key Features

• 2-Step Publishing Flow – “Get trends → Create draft → Schedule → Auto publish”
• Reactive Automation — Keyword → Auto DM / Reply (5-min loop)
• Comment-based Generation – Comments → New draft via adapter
• Playbook Analytics Dashboard – Real-time KPI ↔ Trend correlation
• Graph Explorer UI – Interactive knowledge graph search

🔬 Status

• Fully functional prototype
• Backend ↔ Frontend integration
• Graph RAG sidecar operational
• Threads & Instagram adapters live
• Playbook Analytics dashboard ready 🎥 Demo:https://youtu.be/9QpCtfeVI_o

💡 Vision

This is not just a SaaS product – it’s a blueprint for self-regenerating systems. By demonstrating that knowledge recombination can replace feature addition, Maestro illustrates how future AI platforms will grow intelligence, not code.