From theory to practice.

01Story

The story behind our name

Praxis means turning theory into practice. For us, that means math written in a code editor becoming a working robot. The name Praxtify comes from this idea.

1. THEORY

   Math and code

   From the very start, we define the robot's behavior with equations and software.

2. SIMULATION

   Physics testing

   We try the design step by step in a physics engine and validate it with measurable results.

3. PRAXIS

   A working robot

   We build the design that passes testing under our own brand, then take it to the field.

02Principle

Our principle

Prove it in simulation first, then take it to the field.

This principle connects all our work. We handle every stage ourselves, from design to production and from software to the control panel.

• FOCUS

  What we do

  We build physics-tested robots end to end under our own brand, from mechanical design and electronics to the software and the control panel.

• DIRECTION

  Our goal

  To become the manufacturer people think of first and trust for autonomous robots, starting with living spaces and the everyday tasks within them.

03Toolkit

Our toolkit

We develop our robots end to end ourselves, from mechanical design and electronics to physics testing and AI. These are the tools and technologies we use:

• PHYSICS ENGINE

  MuJoCo

  We test the robot's design and motion in physics-based simulation before it goes to production.

• DEVELOPMENT

  Python

  We write control algorithms, test scenarios and data analysis tools, all of them in Python.

• ROBOT SOFTWARE

  ROS 2

  We build the robot software's components and the communication between them with ROS 2.

• MECHANICAL DESIGN

  CAD

  We draw the robot's parts and assembly as precise technical models, ready for production.

• ELECTRONICS DESIGN

  PCB

  We design the robot's circuit boards ourselves to suit its motor drivers and its sensors.

• EMBEDDED SOFTWARE

  C/C++

  We drive motors and sensors with fast, close-to-hardware code that runs in real time.

• ROBOT COMPUTER

  SBC

  We run the robot's software and AI models on board, using compact single-board computers.

• HARDWARE

  Motors and sensors

  We select motors, drivers and sensors for motion and perception, then integrate them.

• COMPUTER VISION

  OpenCV

  We process camera images so that the robot can perceive its surroundings and objects.

• MACHINE LEARNING

  PyTorch

  We train the robot's perception and behavior models with simulation and real-world data.

• LANGUAGE MODELS

  LLM

  We integrate language models so that people can talk to the robot in natural language.

• VISUALIZATION

  Blender

  We prepare design visuals, promotional content and simulation models for the robot in Blender.