Pashi is the
operating system for manufacturing. It unifies the
entire production process for any product, encompassing
operator instruction & data input interfaces
stage logic & parameter
interfacing & configuration
robot programming &
into a single Pashi Program.
Click on a production process below to watch Aditya Balaraman, a production engineer with experience at Renault-Nissan, Maruti, and Modine, program the logic for it in Pashi
Pashi Programs can be created and modified by both non-programming industrial and production engineers via an intuitive browser-based production-flow GUI and by software developers writing code in the Go programming language. Every change made to a Pashi Program can be tracked using our version control interface, which allows for process modification testing and rollback.
Pashi automatically logs every event and changes in state for each variable in each execution of the Pashi Program at the compiler level, in the built-in SQL-based Pashi Datastore. In combination with our completely customizable report generation & analytics tools, this allows for comprehensive process transparency, execution traceability, and KPI tracking, all of which are ensured even if rapid changes are made to the production process via the Pashi Program.
Pashi gives you the ability to program the stage-to-stage process flow of your production line end-to-end using conditional statements based on arbitrary prcoess paramters and manual input. You can send products to any physically possible stage on the production line, either directly via conveyor or AGV or indirectly by issuing instructions to human operators.
Pashi allows users to create custom screens and check-sheets using our visual human interface creation tool. These forms can be further customized by software developers using HTML-based templating, and can be sent to any human interface device connected to the Pashi Platform, from Android tablets to specialized HMI systems.
Pashi automatically logs every event in each execution of the Pashi Program, including material addition, human & robot actions, machine parameters, and changes in state for each variable, in the built-in SQL-based Pashi Datastore. This allows for complete process and execution traceability.
The data in the Pashi Datastore can be surfaced using our visually customizable report generation tool to produce custom genealogy reports and traceability logs, or analysed using the graphical analytics tool, which allows the entire corpus of production data to be queried, graphed, and compared by non-programmers employing easy-to-use Excel-style formulae. These formulae can also be used to create live, browser-based, permission-controlled dashboards to track important variables and Key Performance Indicators (KPIs).
Pashi allows programmers to write, modify, and extend process control code using a unified programming model in the modern, easy-to-use Go programming language in the text-based environment they are most comfortable with, avoiding the need to write dozens of separate programs using outdated abstractions like Ladder Logic or custom graphical-only activity configuration tools, while still being able to collaborate seamlessly with non-programming industrial engineers.
Powered by the breakthrough research work of computer scientists and roboticists from Yale and Cornell, Pashi also makes the addition of new automation techniques and the coordination of complex interactions between robotic systems -- for example, handoffs of parts from an AGV directly to a robot arm -- simple to implement and scale to tens of thousands of robotic agents using ready-to-use interoperable software modules that can be added to any Pashi Program.
Each Pashi Program, including the associated analytics formulae, dashboards, and report templates, is version-controlled using a Git-based change-request system that allows users to propose modifications to some or all of it by making those modifications on a copy of the program. When a user with the appropriate permissions approves the change, the new version is given a unique identifier. Administrators can change which version of the Pashi Program is running on a given production line with a single click. Each event in the Pashi Datastore is associated with the unique identifier of the Pashi Program version that produced it, allowing for changes to KPIs such as yield and throughput resulting from process changes to be precisely determined in comparison with previous versions of the production process. This allows engineers to roll back erroneous or failed process changes as soon as issues occur, as opposed to at a latency of weeks to months.