how to deploy and serve prodigy ner model as flask api

[Vishal112]

Please help me out I have created a model using prodigy NER now how to serve this as a flask API I have to host it to my internal server and API I have to serve this to frond end pls hence regarding it

[ryanwesslen]

hi @Vishal112,

Thanks for your question.

@Vishal112

I have created a model using prodigy NER now how to serve this as a flask API I have to host it to my internal server and API I have to serve this to frond end pls hence regarding it

If you have created a trained spaCy NER model, then you can treat it as if you'd serve any spaCy model through a Flask API. I'd recommend searching online for resources like this (there are several others):

Building a Flask API to Automatically Extract Named Entities Using SpaCy -...

The key difference is instead of loading a pre-trained pipeline model (e.g., en_core_web_sm) you would load in your NER model (pipeline) you trained. Ideally, you can create this as a python package (spacy package) if you formulate your project as a spaCy project. See this template for how to create Prodigy-integrated spaCy project.

Also, have you considered using FastAPI and/or streamlit instead to serve your model? Both work much easier with spaCy projects and you can see integrations with the same documentation.