The App Engine for Physical World Intelligence
AI that thinks, learns and acts in the physical world
Twinit powers AI-native digital twins of buildings, infrastructure, plants and cities – living semantic graphs integrating IT, OT and engineering data. Custom AI agents reason over the graph; humans in the loop decide.
Three Intelligence Categories
One platform. Three categories of intelligence.
All three share the same composable foundation, semantic graph and AI-native runtime. A given pursuit may involve any combination of them.
PROJECT INTELLIGENCE
Plan, design, build, handover.
Partner-built apps across the capital project lifecycle. Every phase, every use case. The bridge from project to operations is the Asset Twin, delivering the as-built record.
-
Phase-bounded apps: planning, design, construction, and handover
-
One view across schedule, contracts, cost, financials, quality, safety, risk
-
Asset Twin: the as-built bridge into handover
ASSET INTELLIGENCE
The horizontal scaffold.
Scaffolded apps for asset operations. Every asset, every AI capability. The anchor is the Asset Intelligence App Template, the one horizontal scaffold for all industries.
-
AI capabilities: predictive maintenance, health, anomaly, benchmarking, risk-based scheduling, energy
-
Control tower: one view across OT/IoT, CMMS, ERP, contracts, and financials
-
SI-delivered: built on the scaffold by APM specialists
OPERATIONS INTELLIGENCE
Reason across the business.
Bespoke apps for operations outcomes. Every domain, every business question. The anchor is the semantic graph, reasoning across data that no single system holds.
-
Business outcomes: margin, throughput, service levels, compliance, risk
-
Cross-domain reasoning: orders, schedules, contracts, regulations, asset condition
-
Bespoke delivery: per engagement, on the composable platform
The Three-Layer Architecture
Services. Workspaces. Studio.
Studio is where apps are built. Services are what they're built from. Workspaces are where they run. One coherent platform, three concerns separated cleanly.
The platform is unopinionated, composable, and flexible – purpose-built for bespoke apps, not for boxing them in. Developers write scripts; Twinit executes them securely and consistently, insulating applications from infrastructure change, security risk, and scaling concerns.
Twinit Studio
THE DEVELOPER EXPERIENCE
Twinit’s extension for Cursor and VS Code.
Twinit Workbench. App Templates. Open APIs & SDKs.
Twinit Workspaces
THE SECURE RUNTIME
Data & tenancy controls. Secrets and policies. Application Lifecycle.
Sandbox & production isolation. Audit trails.
Twinit Services
THE COMPOSABLE BUILDING BLOCKS
Semantic graph layers. Workflows. AI orchestration. IT and OT / IOT integration. Engineering data integration and visualisation. File management. Evening and Notifications. Custom APIs. MCP server. Authentication. Authorisation. Audit
What makes Twinit AI-Native
AI Is Not Bolted On. It Is the Core Engine.
Six pillars of intelligence, automation, and innovation for apps composed on Twinit.
Twinit is AI-native at its core: Multi-agent orchestration, MCP, graph-aware reasoning, open model choice, RAG grounding, and natural-language interaction, all wired into one platform.
Orchestration
Multi-Agent Orchestration
Compose and govern teams of AI agents that collaborate to detect anomalies, run simulations, and optimise outcomes.
Reasoning
Graph-Aware Reasoning
Semantic knowledge graphs so AI understands context, relationships, and rules – delivering explainable, trustworthy insights.
Authoring
Twinit's extension for Cursor and VS Code
Turn natural language into complete Twinit apps: back-end services, agents, and UI scaffolded rapidly.
Sovereignty
Open Model Choice
Bring your own LLM API keys. Retain sovereignty, flexibility, and control over the intelligence layer.
Grounding
AI Grounding for Context
RAG over your documents, reasoning over the semantic graph, tools to reach into live systems. Agents that answer from your data.
Interface
Natural-Language Interface
Build custom tools that let users interact with apps in natural language. Chatbot-style UX over technical and proprietary data.
Physical World Intelligence, not Physical AI
Two related but distinct domains. Twinit makes the difference explicit because it changes what you build, how you build it, and what you build it with.
Physical World Intelligence
AI that reasons across the project and operational estate.
The physical world is the subject of the intelligence. Agents read graphs, simulate outcomes, recommend actions, and adapt as conditions change.
Physical AI
AI embodied in robots, autonomous vehicles, drones, and humanoids – perceiving and acting in the world directly. Twinit complements it; it doesn't compete with it.
Domain-Agnostic By Design. Vertical-Deep by Partner.
Twinit is horizintal and unopinionated. Partners bring the vertical depth. Every sector across four categories.
Composable Platform. Your Apps. Your IP.
Twinit is the platform you build on. The composable services, the secure runtime, the Al-powered developer experience, the open-source App Templates that compress months of foundational work - all maintained and supported by Twinit. The apps you build for your customers are your IP, your customer relationship, your services revenue.
Building a custom physical-world intelligence app from scratch - semantic graph, telemetry pipeline, Al agent framework, security, lifecycle management, role-based access - is a long undertaking before any domain logic is written. On Twinit it is much faster: Twinit's extension for VS Code and Cursor - the Al-native IDE - supercharges development by combining top-tier models like Claude and GPT-4o with Twinit's specialised logic.
You can build complete Twinit apps using nothing but natural language prompts. The same composition that powers the first use case is the foundation for every subsequent one.