59 integrations
Roboflow Integrations
Explore how to use Roboflow in combination with other software to solve computer vision problems.
/
Annotation and Data
Bring images and labels in from the tools your team already uses.
AWS S3You can add upload images stored in S3 for use in building datasets for models in Roboflow.
CVATPre-label images in CVAT with a Roboflow Universe model, then bring the finished labels into Roboflow to train and deploy your own model.
FastdupDownload a Roboflow dataset, run fastdup to find duplicates, outliers and other problem images, then upload the cleaned set back to train on.
LabelImgUpload the Pascal VOC XML or YOLO files LabelImg saves to Roboflow, then train and deploy a model on your labels.
LabelMeUpload LabelMe JSON files with your images to Roboflow to train detection and segmentation models on your polygons and boxes.
LabelboxUpload a Labelbox JSON export to Roboflow to convert it, train a model on it and deploy the model.
Make SenseConnect a Roboflow model to Make Sense so it draws suggested boxes as you open each image, then upload the finished labels to Roboflow.
SageMaker Ground TruthUpload the output manifest from a SageMaker Ground Truth labeling job to Roboflow, then train, convert or relabel the dataset without a conversion script.
Scale AIUpload the JSON that Scale AI returns for completed annotation tasks to Roboflow, then train and deploy a model on those labels.
SuperAnnotateBring images and labels from SuperAnnotate into Roboflow, then train a model on them and deploy it to the cloud or your own hardware.
SuperviselyExport a Supervisely project and upload it to Roboflow to train, compare and deploy models on the same labels.
VoTTYou can export the data annotated in VoTT into Roboflow for use in generating a dataset with preprocessing and augmentations, and for use in model training.
Training
Train on your Roboflow datasets in the notebook or framework you prefer.
AWS Rekognition Custom LabelsLabel and version a dataset in Roboflow, then send it to Amazon S3 as a SageMaker Ground Truth manifest that Rekognition Custom Labels can train on.
Amazon SageMakerPull Roboflow dataset versions into SageMaker notebooks with the Python package, train there, then upload the weights to Roboflow or serve the model on a SageMaker endpoint.
Amazon SageMaker Studio LabsRoboflow has produced notebooks showing how to train computer vision models that work in Amazon SageMaker Studio Lab.
Azure Custom VisionExport a Roboflow dataset version straight into an Azure Custom Vision project and train it there. Enterprise plans only; Microsoft retires Custom Vision in 2028.
Create MLExport a Roboflow dataset as CreateML JSON and train an object detector in Apple's Create ML app on your Mac, for Core ML models that run on iPhone and iPad.
Google ColabTrain RF-DETR, YOLO26, YOLO11 and other models on your Roboflow datasets in our Colab notebooks, then upload the weights to Roboflow to deploy them.
Hugging FaceImport Hugging Face Hub datasets into Roboflow, and load RF-DETR models, including ones you train on Roboflow, with the Hugging Face Transformers library.
YOLOv5Export Roboflow datasets in YOLOv5 PyTorch TXT, train YOLOv5 in our notebooks, then upload the weights to serve the model from Roboflow's API or your own hardware.
YOLOv8Export Roboflow datasets for YOLOv8, train detection, segmentation or keypoint models with Ultralytics, then upload the weights to run them on Roboflow.
Deployment
Run your model on devices, in the cloud or in the browser.
Amazon EC2Run Roboflow models and Workflows on an EC2 instance in your own AWS account, with one Inference CLI command to provision a GPU or CPU server.
AndroidDeployment
Glitch.meYou can deploy Roboflow models in the browser in a Glitch.me environment.
Google Cloud RunRoboflow Inference ships as a Docker image, so you can run it on Google Cloud Run like any other container. Current Roboflow docs do not include a Cloud Run guide.
Google Cloud VMRun Roboflow models and Workflows on a Compute Engine VM in your Google Cloud project, provisioned with one Inference CLI command.
Google Cloud VisionYou can train models on datasets in the Roboflow platform directly on the Google Cloud Platform through Cloud Vision.
IBM CloudYou can export the data annotated in IBM Cloud into Roboflow for use in generating a dataset with preprocessing and augmentations, and for use in model training.
KubernetesRun the Roboflow Inference server in your own Kubernetes cluster, from a single pod and service to an autoscaled deployment with the Enterprise Helm chart.
Luxonis OAKRoboflow provides a Python package called roboflowoak that provides easy utilities for running models hosted on Roboflow on an OAK.
NVIDIA JetsonYou can deploy models hosted on Roboflow to an NVIDIA Jetson.
NVIDIA TRTSpeed up Roboflow models on NVIDIA GPUs and Jetson devices by turning on TensorRT in the Roboflow Inference server.
OpenRouterRun 300+ vision-language models from OpenRouter in Roboflow Workflows with one block, for OCR, captioning, classification, detection and structured answers.
Raspberry PiRun the Roboflow Inference server in Docker on a Raspberry Pi 4 or 5 and serve your models and Workflows from the device.
ReplitStart from the Roboflow Web Quickstart template on Replit to publish a web page that runs your model on a webcam feed in the browser.
Web BrowserRun Roboflow models in a web page with inferencejs, on images or a webcam, or stream video to Roboflow's servers with the web inference SDK.
iOSRun your Roboflow model on iPhone and iPad with the Roboflow Mobile iOS SDK, for on-device object detection and instance segmentation in native apps.
Connections
Read and write PLC tags, and send results to SCADA, databases and any API.
AVEVA InTouch and System PlatformShow detections, counts and pass or fail results in AVEVA InTouch and System Platform as live tags, served from Roboflow's OPC UA Server on your edge device.- AVEVA PI SystemStore detections, counts and pass or fail results as PI points, collected from your edge device over OPC UA with the AVEVA Adapter or PI Connector for OPC UA.
Allen-Bradley PLCsRead and write ControlLogix, CompactLogix and Micro800 tags over EtherNet/IP, so a detection can reject a part, hold a station or count output.
ChatGPTFind datasets, label images, train models and build Workflows in Roboflow by asking ChatGPT, with the Roboflow plugin connected to your workspace.
ClaudeManage datasets, train models, build Workflows and run inference in Roboflow by asking Claude, with Roboflow connected from the Claude directory.
Claude CodeManage datasets, train models, build Workflows and run inference in Roboflow from Claude Code, with the Roboflow plugin and MCP server.- CodexRun inference, build Workflows and manage datasets and models in Roboflow from Codex, with the Roboflow plugin and MCP server.
CursorManage datasets, train models, build Workflows and run inference in Roboflow from Cursor, with the Roboflow plugin and MCP server.
IgnitionWrite detections, counts and pass or fail results from Roboflow Workflows into Ignition tags over OPC UA or MQTT, for Perspective screens, history and alarms.
Kepware ServerWrite detections, counts and pass or fail results into Kepware Server tags over OPC UA, so every device and client connected to Kepware can use them.
Matrikon OPCWrite detections, counts and pass or fail results to MatrikonOPC servers through Matrikon OPC UA Tunneller, so the devices and clients behind them can use them.
Microsoft SQL ServerInsert detections, counts and pass or fail results from Roboflow Workflows as rows in a Microsoft SQL Server table, ready for reports, dashboards and other systems.
Mitsubishi MELSEC PLCsRead and write MELSEC iQ-F devices over Modbus TCP, so a detection can reject a part, hold a station or count output on FX5 PLCs.
Modbus TCPRead and write coils and registers on any Modbus TCP device, so a detection can reject a part, signal a robot or count output.
OBSRun a Roboflow model on your OBS Studio output in the browser and switch scenes or move sources when it detects a gesture.
Omron NX/NJ controllersRead and write Omron NX and NJ controller values over Modbus TCP, so a detection can reject a part, hold a station or count output.- Process HistoriansLog detections, counts and pass or fail results in AVEVA PI, Canary, FactoryTalk Historian and other historians, collected over OPC UA from your edge device.
- Robot controllersSend pick, sort and reject signals to Universal Robots and FANUC controllers over Modbus TCP, and read their status back into a Workflow.
SalesforceTurn a Workflow detection into a Salesforce Work Order or Case with the evidence image attached, using one custom block you configure in your browser.
Siemens S7 PLCsRead and write S7-300, S7-400, S7-1200 and S7-1500 data through PLC Relay, so a detection can reject a part, hold a station or count output.
SlackPost a Slack message, with the annotated image attached, whenever a Roboflow Workflow finds something, using the Slack Notification block.
ZapierRun Roboflow models inside Zaps with the Roboflow app on Zapier, or send Workflow results to a Zapier webhook to act on them in thousands of apps.
Don’t see your tool?
Roboflow imports and exports the common annotation formats and has a REST API and Python SDK, so most stacks connect without a dedicated integration.