Visual Intelligence Summit: Oct 22 in San Francisco Get your ticket (opens in a new tab)

Roboflow + Robot controllers

Send pick, sort and reject signals to Universal Robots and FANUC controllers over Modbus TCP, and read their status back into a Workflow.

Many robot controllers can act as a Modbus TCP server, exposing general-purpose registers that a robot program reads and sets. Universal Robots controllers run one by default on port 502; FANUC controllers add one with a Modbus TCP software option.

Roboflow's PLC Writer block writes a Workflow result, such as a class code or a pick position, to those registers, and PLC Reader reads values back, such as a robot-ready bit. The blocks connect directly or through PLC Relay, a service on your edge device. The robot program decides what to do with each value, so motion stays under the controller's control.

How to connect Robot controllers to Roboflow

6 steps
  1. Set up an edge device

    Register a device in Deployment Manager and connect it to the same network as the robot controller, with port 502 reachable. The device runs your Workflows on site.

  2. Enable the robot's Modbus TCP server

    On a Universal Robots controller the Modbus TCP server is already running. On a FANUC controller, install the Modbus TCP software option and enable the server.

  3. Pick the registers to use

    From the robot's Modbus register map, choose general-purpose registers for the values the Workflow will write, such as a class code, and the ones it will read, such as a ready bit.

  4. Add the PLC Relay and define tags

    On the edge device's Configuration tab, add the PLC Relay service, choose Modbus TCP and enter the robot controller's IP address. Add each register as a tag by its address from the robot's register map, such as holding:100, and mark the ones the Workflow may write.

  5. Add PLC blocks to the Workflow

    Add a PLC Reader to bring robot status in and one PLC Writer for each value you send, pointed at the relay. To skip the relay, set Connection mode to Direct (Modbus TCP) and enter the controller's IP address and port.

  6. Use the values in the robot program

    In the robot program, read the register and branch on its value, for example a switch on the class code that picks the right bin.

What teams build with Robot controllers

  • Sort parts by class

    Write a class code for each part so the robot program places it in the right bin or lane.

  • Remove a defective part

    Write a reject value when the Workflow finds a defect, and let the robot program pick the part off the line.

  • Wait for the robot to be ready

    Read a ready register before the next inspection, so the Workflow only sends a result when the robot can act on it.

Works with

ProductHow Roboflow connects
Universal RobotsBuilt-in Modbus TCP server
FANUCModbus TCP software option

Questions

Any robot controller that acts as a Modbus TCP server. Universal Robots controllers have one built in, and FANUC controllers need the Modbus TCP software option. For other brands, check the controller's documentation for Modbus TCP server support.

No. Roboflow writes values to the robot's registers, and the robot program decides how the robot moves in response.

The direct EtherNet/IP mode of the PLC blocks reads and writes Allen-Bradley Logix tags. For robot controllers, use Modbus TCP.

No. The PLC blocks run on your own device on the plant network, close to the camera and the robot, not on Roboflow's hosted API.

The PLC blocks and PLC Relay are part of the Enterprise plan. Talk to our team to add them to your workspace.

Why are companies choosing Roboflow?

The end-to-end platform for vision AI: create datasets, train models and deploy to production.

  • Flexible deployment modes

    Move from a scalable cloud API to edge hardware like NVIDIA Jetson without managing complex dependencies.

  • Integrated workflow builder

    Combine multiple AI models, custom business logic and third-party integrations into one production-ready pipeline in a low-code editor.

  • Model monitoring and alerts

    Dashboards track confidence, inference time and request volume for your deployed models, and email alerts tell your team when confidence suddenly drops.

View all
Get started

Build on the Platform

For developers, engineers, and technical founders who want to get hands on. Try the free tier; the docs are open.

Bring it into your operation

For heads of AI, operations leaders, and enterprise teams. Bring a known problem, or work with us to find the first one worth taking on.