EyePick provides Maestro, an industrial controller that consolidates camera, robot, sensor, and PLC connectivity with an integrated computer‑vision stack and graph‑based AI decision engine. The platform lets manufacturers and logistics operators create, deploy, and manage vision‑driven inspection, sorting, and palletizing workflows as reusable recipes, reducing integration effort and supporting plug‑and‑play hardware across multiple production cells.
Funding
Funding not disclosed
Founders
Product
Problem
Manufacturers and logistics operators often face fragmented integration of cameras, robots, sensors, and PLCs, requiring custom drivers, manual orchestration, and costly maintenance for AI‑driven inspection, sorting, and palletizing processes. This complexity hampers rapid deployment of computer‑vision models and limits scalability across production cells.
Solution
EyePick delivers Maestro, an all‑in‑one industrial controller that unifies hardware connectivity, computer‑vision pipelines, and AI decision logic on a single platform. The system abstracts device drivers for cameras, robots, grippers, and PLCs, allowing users to map signals once and reuse them across multiple cells. Built‑in vision modules support 2D, 3D, SWIR, and hyperspectral imaging, while a graphical decision engine lets engineers construct AI‑driven workflow graphs with guardrails, timeouts, and PLC alignment. Applications are packaged as “recipes” that can be deployed, monitored, and rolled back through a centralized console, reducing integration effort and accelerating time‑to‑value for physical AI use cases.
Target Audience
Primary customers are manufacturing and logistics firms that require automated visual inspection, sorting, bin‑picking, and palletization, as well as system integrators building AI‑enabled production lines.
Features
- Unified hardware abstraction layer with maintained drivers for cameras, robots, grippers, conveyors, sensors, and PLCs, enabling plug‑and‑play connectivity
- Integrated vision stack covering 2D, 3D, profilometry, SWIR, and hyperspectral imaging, pre‑configured for common inspection and sorting tasks
- Graph‑based AI decision engine that links vision outputs to PLC logic, supporting conditional branching, guardrails, and timeout handling
- Simulation and debugging environment that replicates the production cell, allowing offline testing and rapid iteration of vision‑decision pipelines
- Centralized logging, trace collection, and version‑controlled recipe management for easy rollback and auditability
- Flexible deployment via selectable communication protocols (e.g., OPC UA, MQTT, EtherNet/IP) and customizable dataframes for downstream systems
- Web‑based Maestro Studio interface for orchestrating, monitoring, and maintaining multiple cells from a single dashboard
- Compatibility layer that augments existing PLCs rather than replacing them, preserving legacy control investments