# Prompt flow

A suite of tools for building LLM applications: create flows that link prompts, LLMs and Python code, then test, evaluate and deploy them to production.

Prompt flow is a suite of development tools for building applications powered by large language models. It covers the full cycle from ideation and prototyping to testing, evaluation, and production deployment and monitoring, and is meant to make prompt engineering easier while keeping production quality.

A flow is created by linking large language models, prompts, Python code and other tools together into an executable pipeline; a command-line tool tests a flow interactively and manages connections such as Azure OpenAI credentials. Flows can be evaluated against larger datasets, and that evaluation can run inside a CI/CD pipeline. A finished flow deploys to a serving platform or integrates directly into an application's own code.

An optional cloud version in Azure AI adds team collaboration on top of the same flows.

Website: https://microsoft.github.io/promptflow/