> This page location: Docs > Reference > Overview > Full index for agents: https://www.semantius.com/llms.txt # Overview > What is the Semantius Agentic Data Platform. - **URL**: https://www.semantius.com/docs/overview --- Semantius is an **agentic data platform**. It lets you describe your business as a semantic model: the entities you care about, the attributes that define them, and the relationships between them. It then puts that model to work for both humans and AI agents. Instead of stitching together a database, an admin UI, an API layer, and a permissions system every time your domain evolves, you evolve the model. Everything else follows. ## Semantic Models At the center of Semantius is the **[semantic model](https://www.semantius.com/docs/models.md)**. A semantic model captures: - **Entities**: the concepts in your domain (customers, orders, projects, devices). - **Fields**: the attributes of each entity, with human-readable labels, descriptions, and UI hints. - **Relationships**: how entities reference or own each other (`reference` for cross-entity links, `parent` for ownership and composition). - **RBAC**: roles, permissions, and the rules that govern who can read or change what. The model is data, not code. That means it can be inspected, edited, versioned, and (crucially) manipulated by agents. ## MCP Servers for AI Chats Semantius exposes its model through **[MCP servers](https://www.semantius.com/docs/mcp-connectors.md)**, so you can connect AI chats like **ChatGPT** and **Claude** directly to your data. From a chat window you can: - Ask questions about your data in natural language and get answers grounded in the semantic model. - Create, read, update, and delete records with the model's permissions enforced. - Evolve the schema by describing the change you want: add entities, fields, and relationships. - Run analytical queries through a semantic layer that understands measures, dimensions, and joins. The MCP layer is the bridge between conversational AI and your operational data, with the semantic model providing the vocabulary. See [MCP Connectors](https://www.semantius.com/docs/mcp-connectors.md) for the full list of servers and clients, and [Installation](https://www.semantius.com/docs/mcp-connectors/installation.md) for setup. ## Agentic Skills Semantius is **agentic** by design. It ships with **[skills](https://www.semantius.com/docs/agent-skills.md)** that agents can use to build, test, and deploy models end to end: - [Generate a model](https://www.semantius.com/docs/models/create.md) from a description of your domain. - Add or refactor entities and fields without breaking existing data. - Configure roles, permissions, and access policies. - [Deploy changes](https://www.semantius.com/docs/models/deploy.md) to a live environment, then [optimize](https://www.semantius.com/docs/models/optimize.md) by pulling the live module back into a fresh spec. Skills give agents a structured, repeatable way to do real work on your data, not just talk about it. Start with [ready-to-use templates](https://www.semantius.com/docs/models/templates.md) or build a model from scratch with the [Business Analyst](https://www.semantius.com/docs/models/create.md) skill. ## CLI for Coding Agents For coding agents like **OpenClaw** and **Claude Code**, Semantius provides a **[CLI](https://www.semantius.com/docs/agent-skills/installation.md)**. The CLI lets agents work with your data autonomously from the terminal: introspect schemas, query data, apply changes, run validations, and integrate the results back into your codebase. This makes Semantius a natural fit for workflows where an agent needs to reason about your domain while writing code, whether that means generating typed clients, building dashboards, scaffolding features, or migrating data. ## Typical Setup After creating a new data platform on [app.semantius.com](https://app.semantius.com), you: 1. **[Set up the MCP servers](https://www.semantius.com/docs/mcp-connectors/installation.md)** so you can chat with your platform from any MCP-compatible client (ChatGPT, Claude, and others). 2. **[Install the CLI and skills](https://www.semantius.com/docs/agent-skills/installation.md)** in any agent that supports Agent Skills, such as OpenClaw or Claude Code. From there, install one of the [ready-to-use models](https://www.semantius.com/docs/models/templates.md) we provide, customize it, or use the **[Business Analyst](https://www.semantius.com/docs/models/create.md)** skill to build a custom model from scratch. The result is a single source of truth that humans, conversational AI, and coding agents can all use, speaking the same semantic language. --- ## Related - [Reference](https://www.semantius.com/docs/reference.md) - [Models](https://www.semantius.com/docs/models.md) - [Docs home](https://www.semantius.com/docs.md) - [Full index for agents](https://www.semantius.com/llms.txt)