Artly provides a generalized training and learning software platform for humanoid robots to execute complex physical tasks. This AGI system uses LLMs to compose goal-driven action sequences, enabling scalable, intelligent manipulation skills from day one. The platform supports data collection, behavior optimization, and cloud-based deployment of foundation models for commercial robotics applications.
Funding
Funding not disclosed
Founders
Product
Problem
Traditional robot programming is complex, slow, and requires specialized robotics engineers, creating a barrier to entry for businesses needing automation solutions. Existing systems often struggle to adapt to new environments and tasks, limiting their flexibility and increasing operational costs.
Solution
Artly AI offers a software platform that enables robots to learn complex manipulation tasks through human demonstration, streamlining the training process and reducing the need for specialized programming skills. The Artly General Intelligence (AGI) Platform allows robots to perform tasks with human-level dexterity and adaptability, using a combination of imitation learning and AI-driven task composition. By leveraging a large language model (LLM), the platform can break down high-level instructions into structured steps, enabling robots to execute complex, multi-step tasks with minimal human input. This approach allows for the creation of general-purpose manipulation robots capable of grasping, picking, cutting, pouring, and stirring, with applications ranging from barista services to food preparation.
Target Audience
The primary target audience includes businesses in the food and beverage industry, and other sectors requiring automation of complex manipulation tasks, as well as robotics researchers and developers seeking a versatile AI platform for robot control.
Features
- AGI Platform: A modular system for training, refining, and scaling robotic intelligence.
- Imitation Learning: Robots learn from human demonstrations, enabling fast and intuitive programming.
- Task Composition: LLM-driven system breaks down high-level instructions into structured steps for robot execution.
- Skill Library: Growing list of pre-trained skills, including grasping, picking, cutting, pouring, and stirring.
- AGI Hub: Cloud-based infrastructure for training, sharing, and deploying robot foundation models; supports various robotic arms and hands.
- Model Marketplace: Users can publish and monetize their own robot models.
- Vision-based Control: Robots adapt their grip using tactile sensors and vision-based control.