Intelligence Layer PLEXA Core 2.0 | AI-Agents

PleXA CORE 2.0

The layer on which the agents collaborate.

In the Intelligence Layer, specialized AI agents work together, trained on data from thousands of real-world applications. Using VLA models, they process real-world input from vision and sensors, as well as user input via speech, to understand processes, plan tasks, and make decisions. Through integration with horstOS, these decisions are directly translated into actions. This allows the robot to react flexibly to its environment and perform tasks independently. This is how Physical AI becomes practical: simple, flexible automation—without complex programming.

AUTOMATION AGENTS

Every agent is an expert.

Specialized agents handle the planning, implementation, and operation of your application. Depending on the task at hand, the appropriate agents are specifically selected to contribute their expertise and help achieve the best possible result. All agents always operate within clearly defined roles, permissions, and authorizations, ensuring that you, as the user, never lose control over your processes.

AGENT

Project Creation

Creates setups from voice & image inputs.

Learn more

AGENT

Logic
Builder

Creates robot logic and programs.

Learn more

AGENT

Safety &
Error Handling

Monitors operations and prevents errors.

Learn more

AGENT

Operations & Support

Assists users with their questions.

Learn more

AGENT

Supervision & Execution

Coordinates agents and controls the robot.

Learn more

| PROJECT CREATION AGENT

Project Setups Created in Seconds

The Project Creation Agent generates automation concepts based on inputs such as process descriptions, CAD data, or images. This enables the feasibility of an automation request to be assessed in a very short time and a suitable concept to be provided.

| Logic Builder Agent

Program Logic at the Push of a Button

The Logic Builder Agent creates program workflows based on described process objectives or predefined videos. During operation, it continuously optimizes the workflows and adapts them to changing conditions.

| Safety & Error Handling Agent

Everything at a glance, anytime

The Safety & Monitoring Agent acts as an independent safety authority and monitors all operations in real time. It detects potential risks—such as collisions or impending failures—early on, before they occur. When problems arise, it analyzes the cause and provides troubleshooting guidance to minimize downtime.

| Operations & Support Agent

Always by your side

The Operation & Support Agent supports day-to-day operations by customizing programs and continuously optimizing processes. In addition, the agent serves as an active assistant to users, answering questions, providing recommendations, and helping to ensure that robot deployment runs efficiently and smoothly.

| Supervising and Execution Agent

Everything's under control

The Supervising & Execution Agent is the central coordinator in the Intelligence Layer. It monitors and orchestrates all specialized agents, continuously processes their input, and synchronizes it with data from vision and sensors in the real world. Based on this, it derives specific actions in real time and translates them directly into control commands. It is the agent that actually moves the robot in the real world.