Tool in Foundry and Azure Machine Learning for assembling LLM applications as graphs of nodes (LLM, prompt and Python tools), with variants and evaluation. Microsoft Agent Framework is replacing it as it is retired.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Prompt flow in context, with comparison tables and the common traps.
Terms in this definition
- Chat message roles
Labels on chat messages: instructions go under system, the person's input under user, the model's previous answers under assistant, and results returned by a called tool under tool (or function).
- Azure Machine Learning
Azure platform where you train, deploy, monitor and retrain models of your own, practising MLOps through tools such as pipelines, online endpoints, model monitoring and prompt flow.
- LLM
Large language model, usually a transformer network with billions of parameters that learnt to predict the next token from vast amounts of text. Compared with a small language model it is more capable across tasks, but slower and costlier.
- Microsoft Agent Framework
Open-source SDK that succeeds AutoGen and Semantic Kernel. It is used to build agents and graph-based multi-agent workflows, with support for tools, memory, session state and human-in-the-loop steps.
Related terms
- PF_DISABLE_TRACING
Environment variable governing tracing in prompt flow, which is disabled unless changed. Setting it to false reveals the Trace tab, showing duration and token cost for each node.
- Span
One operation inside a trace, such as a single LLM call or prompt flow node, carrying attributes plus start and end times. Nested spans reveal the order of calls.
- start_trace
Called from the prompt flow SDK, it sends traces to a local trace server and UI, unless export has been configured. By itself it will not capture traces into a Foundry project.