Real-time, physics-informed AI platform for predictive drilling operations.

DeepBoreAI is a vendor-agnostic, plug-and-play system that deploys intelligent ML agents to monitor, predict, and mitigate critical drilling issues in real time. Designed for both conventional and unconventional well types, it combines physics-informed machine learning with cutting-edge edge computing to deliver high-precision situational awareness on the drill floor.

• Real-Time Telemetry Ingestion (WITSML-compatible)
• Physics-Informed ML Agents for:
  • Mechanical Sticking
  • Differential Sticking
  • Hole Cleaning
  • Mud Losses / Washouts
  • ROP Optimization
• Self-Learning Models that adapt to drilling conditions
• Edge Processing for ultra-low latency alerts
• Interactive Dashboard with real-time data and trend visualization
• Historical Analytics + CSV Export
• One-Click Deployment with Docker Compose

• Docker + Docker Compose
• Node.js + npm (if running frontend separately)
• Python 3.10+

git clone https://github.com/ttracx/deepboreai.git
cd deepboreai
docker-compose up --build

• Backend API: http://localhost:8000
• Frontend UI: http://localhost:3000

[WITSML Source] -> [FastAPI Backend] <--> [ML Agent Services]
                             |
                             V
                   [WebSocket + REST API]
                             |
                             V
                   [React Frontend Dashboard]

deepboreai/
├── data_ingestion/          # FastAPI backend
├── frontend/                # React.js dashboard
├── Dockerfile.backend
├── frontend/Dockerfile.frontend
├── docker-compose.yml
├── drilling_data.db         # Local SQLite database
└── README.md

© 2025 DeepBoreAI. All rights reserved.