microsoft/agent-lightning

★ 17,176PythonAudience · developerComplexity · 4/5Setup · moderate

Mindmap

mindmap
  root((Agent Lightning))
    What it does
      Agent training
      Zero code changes
      Framework agnostic
    Supported frameworks
      LangChain
      AutoGen
      CrewAI
      OpenAI SDK
    Optimization methods
      Reinforcement learning
      Prompt optimization
      Fine-tuning
    Architecture
      Tracer spans
      LightningStore
      Trainer loop

mindmap root((Agent Lightning)) What it does Agent training Zero code changes Framework agnostic Supported frameworks LangChain AutoGen CrewAI OpenAI SDK Optimization methods Reinforcement learning Prompt optimization Fine-tuning Architecture Tracer spans LightningStore Trainer loop

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Improve the accuracy of a LangChain agent using reinforcement learning without rewriting it.

USE CASE 2

Automatically optimize the system prompt of an OpenAI-SDK agent based on recorded execution traces.

USE CASE 3

Fine-tune a CrewAI multi-agent system on real task examples to reduce failure rate.

USE CASE 4

Record tool calls and rewards from an existing AutoGen agent and use them to train a better policy.

Tech stack

Pythonpip

Getting it running

Difficulty · moderate Time to first run · 1h+

Requires an existing agent framework and a reward function, install via pip as agentlightning.

License not mentioned in the explanation.

In plain English

Agent Lightning is a training toolkit from Microsoft for AI agents. The idea is to take an agent you have already built and improve its behavior using machine-learning techniques, with as close to zero code changes as possible. Rather than forcing you to rewrite your agent, Agent Lightning slips in alongside it and learns from how the agent actually runs. It is framework-agnostic. You can plug in agents built with LangChain, OpenAI Agent SDK, AutoGen, CrewAI, Microsoft Agent Framework, or even agents written directly against the OpenAI API with no framework at all. In a system where several agents work together, you can choose to optimize one, some, or all of them. The supported optimization approaches include Reinforcement Learning, Automatic Prompt Optimization, and Supervised Fine-tuning. Under the hood, your agent keeps running normally. A lightweight tracer, or an explicit emit helper you drop into code, records every prompt, tool call, and reward as structured spans. Those spans flow into a central component called the LightningStore, which keeps tasks, resources, and traces in sync. On the other side, an algorithm reads the spans, learns from them, and writes back updated resources such as refined prompt templates or new policy weights. A Trainer streams data to runners, moves resources between the store and the algorithm, and updates the inference engine when improvements land. You would use this if you have a working agent that you want to make smarter or more accurate without throwing it away and starting over. The package is installed with pip under the name agentlightning, and the codebase is Python.

Copy-paste prompts

Prompt 1

Using Agent Lightning, show me how to add a tracer to a LangChain agent and run reinforcement learning optimization on it.

Prompt 2

How do I install agentlightning and connect it to an existing OpenAI API agent to do automatic prompt optimization?

Prompt 3

Give me an example of defining a reward function in Agent Lightning for a customer-support agent that should minimize response length.

Prompt 4

Show me how to use Agent Lightning's LightningStore to inspect the spans recorded from a multi-agent CrewAI workflow.

Prompt 5

How do I do supervised fine-tuning on an AutoGen agent using Agent Lightning with my own labeled examples?

Open on GitHub → Explain another repo

← microsoft on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.