Phidata

What is Phidata?

Phidata is an open-source tool for building, deploying, and monitoring AI applications. It ships with pre-built templates, runs applications locally with Docker, and deploys to AWS or as Function as a Service (FaaS) for scaling. It also continuously monitors AI apps for quality and performance.

Pre-built templates cover several application types, including AI apps, AI APIs, Django apps, Streamlit apps, and a template for junior data engineers. Each template comes pre-configured with the required components, so applications can be up and running quickly.

Who created Phidata?

Phidata was created to speed up AI product development by providing pre-built templates for popular frameworks such as FastApi, Django, and Streamlit. Its goal is to help developers and their teams build and deploy AI applications faster, with an emphasis on monitoring quality and performance.

Who is Phidata for?

Phidata is aimed at AI engineers and works for both individual developers and teams.

Getting started with Phidata

1. Create your application: start from a pre-built template with phi ws create.
2. Test locally: run phi ws up to verify everything works as expected.
3. Deploy to AWS: once local testing passes, deploy with phi ws up prd:aws.

Pick the template that fits your project—a general AI app or a Django app, for example—and start building. Templates are production-ready, which reduces configuration work before deployment. Once running, Phidata’s monitoring helps you fine-tune quality and performance over time.