Roboflow + Azure Custom Vision
Export a Roboflow dataset version straight into an Azure Custom Vision project and train it there. Enterprise plans only; Microsoft retires Custom Vision in 2028.
How to connect Azure Custom Vision to Roboflow
5 stepsCopy your Custom Vision key and endpoint
In Azure, open your Custom Vision project and its API key page. Copy KEY 1 and the Endpoint.
Add them to Roboflow
In your Roboflow workspace settings, paste KEY 1 into the training-key field and the endpoint into the Endpoint field. You only do this once.
Generate a dataset version
Create a version with the preprocessing and augmentation you want Azure to train on.
Export as Azure Custom Vision
Choose Azure Custom Vision as the export format. Stay signed in to Azure in the same browser, and the version appears in your Custom Vision account.
Train in Custom Vision
Train the uploaded project in the Custom Vision portal. Azure requires the S0 pricing tier for datasets over 1,000 images.
What teams build with Azure Custom Vision
Keep an Azure Custom Vision pipeline running
Label and augment data in Roboflow while your existing Custom Vision models stay in production until you migrate.
Compare Custom Vision with RF-DETR
Train Custom Vision and a Roboflow model on the same dataset version to plan your move off a retiring service.
Questions
Yes. Microsoft will fully support existing Custom Vision customers until September 25, 2028, and recommends planning a transition before then.
Our docs list the Azure Custom Vision training integration as available on Enterprise plans only.
For datasets larger than 1,000 images, Azure requires the S0 pricing tier on your Custom Vision resource.
Roboflow is listed on the Azure Marketplace, and customers with a Microsoft Azure Consumption Commitment can count the purchase toward it.
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.
