Agent Lightning v1.0.1 marks the first official release of the Agent Lightning Skill, which helps coding agents optimize other AI agents.

Provide an editable agent and a benchmark, and the skill guides systematic improvements to prompts, tools, workflows, models, and reasoning settings—balancing accuracy, cost, latency, and reliability through measured iteration.

Install it for Claude Code, Codex, or GitHub Copilot:

gh skill install microsoft/agent-lightning agent-lightning --agent <agent>

This release also strengthens CI, packaging, release automation, documentation, and benchmark reporting.

Agent Lightning v1.0.0 is released! In this release, we have completely refactored the codebase, making it lightweight while adding support for the latest models. Agent Lightning v1.0.0 is designed for RL training with the exact agent harness used in deployment and addresses several key challenges in this paradigm.

Hightlights

• 🪶 ~3,500 lines of code: We treat simplicity as the first principle.
• 🧩 Train with real agent harnesses: Agents interact with the model through the Agent Lightning v1.0 proxy with ZERO changes, while keeping tools, context, control flow, and environments in the loop.
• ☸️ Native Kubernetes support: Run agents directly as Kubernetes Jobs without relying on external sandbox services.
• 💻 Full coding agent training example: Using only 6K training samples, an end-to-end Qwen3.5-9B workflow improves SWE-bench Verified from 41.8% to 56.4%, a gain of 14.6 percentage points. We release the full pipeline, including data cleaning, reward-hacking prevention, and training scripts.

Agent-lightning v0.3.0 is a major release that introduces several new features and bug fixes. The release is a collaborative effort between Agent-lightning core teams and the community. Thanks to all the contributors who made this release possible.

Highlights

• Tinker integration: Support Tinker as an alternative backend for Reinforcement Learning (#226 #245 #264 #269 #327). See example code, blog 1 and blog 2.
• Azure OpenAI integration: Support Azure OpenAI as a backend for LLM inference and supervised fine-tuning (#256 #327). Example code.
• MongoDB-based Lightning Store is added as an alternative backend for Lightning Store (#323). Documentation.
• Contrib package: Add contrib package for community projects. Search-R1 is integrated as a contrib recipe. More coming. (#239 #396 #410 #412 #417).
• RESTful API: Stabilize and document RESTful API for Lightning Store (#241 #275). Documentation.
• OTel Semantic Conventions that are specifically designed for Agent-optimization areas (#340). Documentation.
• [Preview] Agent-lightning Dashboard is now available (#288 #289 #291 #296 #371 #375). It's the official web application for inspecting and debugging Agent-lightning experiments. See details here.
• [Preview] Multi-modality example featuring VERL and a LangGraph agent on ChartQA dataset (#379). Example code.
• [Preview] Integrate Claude Code as a LitAgent and support training on SWE-Bench (#332 #346 #348). Example code.
• [Preview] Weave tracer as a substitute for AgentOps tracer (#277 #411 #420 #423). Documentation.
• [Preview] Trajectory Level Aggregation for more efficient training with VERL. See blog and documentation.

Store Benchmark

In this release, the Lightning Store core was redesigned for significantly greater efficiency and scalability (#315 #318 #328 #342 #344 #356 #380 #388 #418 #421). The benchmark results below demonstrate the impact: with large numbers of concurrent runners, v0.3.0 delivers up to a 15x increase in throughput compared to v0.2.2.

Notes:

1. Benchmarks were run on a single Standard_D32as_v4 Azure VM (Large and heavy trace tests used Standard_D64ads_v5), executed via GitHub Actions.
2. Two algorithm patterns are evaluated: the batch pattern submits a group of rollouts and waits for all to finish before starting the next group, while the queue pattern maintains a set number of in-flight rollouts, submitting new ones as soon as capacity frees up. Configuration details are available here.
3. The number of turns is directly proportional to the number of spans each rollout generates.

Maintenance and Bug fixes

Core (Store, Interfaces, etc.)

• Add Trainer port option for client-server strategies (#198)
• Fix store port conflict handling (#227)
• Unified PythonServerLauncher (#286 #292 #303)
• Make health timeout configurable (#305)
• Refactor logging (#306)
• Support OTLP in LightningStore (#313)
• Centralized metrics helper (#368)
• Fix redundant cancel tracebacks on Ctrl+C (#370)

Proxy, Adapters and Algorithms

• Fix training metrics before and after processing in VERL (#145)
• Forward streaming requests for Anthropic and OpenAI APIs (as non-streaming requests) (#299)
• Check traces with reward for VERL (#317)
• Patch LiteLLM root span (#341)
• Handle ref_in_actor flag for LoRA compatibility (#386)
• Support with_llm_proxy and with_store in algorithms (#398)
• Support image URL export in TracerTraceToTriplets (#400)
• Fix match_rewards assign_to elements in TraceTree (#403)
• Support customizing trainer and daemon in VERL (#407)

Runners, Tracers and Agents

• Refactor tracer initialization (#321)
• Fix OpenAI Agents 0.6 compatibility (#322)
• emit_operation, emit_annotation, tags and links (#359)
• Sunset HTTP tracer (#402)

Examples

• Fix typos in train-first-agent.md (#263)
• Fix room_selector example which always runs the first task (#270)
• Fix typo in SQL agent example (#285)
• Add the README and script files for training SQL agent on NPU (#272)
• Examples Catalog and Refine Contribution Guide (#331)
• Upgrade LangChain to 1.x (#364)
• Update RAG example to Agent-lightning v0.2.x (#349)

Miscellaneous

• DeepWiki Badge (#263)
• Add AGENTS.md (#374)

New Contributors

Warm welcome to our first-time contributors: Vishal V (@cptnm3), Shenghua Chen (@TerryChan), 부창규 (@genji970), Geng Zhang (@zxgx), xiaochulaoban, Leonardo Pinheiro (@lspinheiro), etsplz (@Kwanghoon-Choi), Vasu (@Vasuk12), Yingluo Li (@totoluo), jinghuan-Chen 🎉

Full Changelog: v0.2.0...v0.3.0

Agent-lightning v0.2.2 is a stabilization release for v0.2.1. It introduces several bug fixes.

• Fix compatibility issues with VERL 0.6.0.
• Fix model name for pre-downloaded models in VERL.
• Fix preparing status transition on rollout when creating attempts.
• Fix OpenAI Agents SDK compatibility issues.

Full Changelog: v0.2.1...v0.2.2

Agent-lightning v0.2.1 is a stabilization release for v0.2.0. It introduces several bug fixes and new features, plus a number of unlisted CI improvements.

Bug fixes

• Fix LiteLLM issues when restarting the proxy multiple times in the same process (#174 #206)
• Fix LiteLLM model name selection when multiple servers use the same model (#197)
• Fix store port conflict handling (#227)

New Features

• Add trainer port option for client-server strategies (#198)

Documentation

• Add tutorial for launching workers on separate machines (#213)
• Add link to VERL framework (#210)
• Add link to vLLM blog (#215)
• Fix a couple of typos and avoid emacs backup files (#237)

New Contributors

A warm welcome to our first-time contributors: scott-vsi, Zhang Zitong (@ddsfda99), John Eismeier (@jeis4wpi) 🎉

Full Changelog: v0.2.0...v0.2.1

Agent-Lightning v0.2.0 introduces major framework improvements, new execution strategies, expanded documentation, and enhanced reliability across the agent training and deployment workflow. This release includes 78 pull requests since v0.1.2.

Core Enhancements

• Lightning Store: Added unified interface and implementation for Agent-lightning's core storage.
• Emitter: Emitting any objects as spans to the store.
• Adapter and Tracer: Adapting to OpenAI-like messages, and OpenTelemetry dummy tracer.
• LLM Proxy: Added LLM Proxy as the first-class citizen in Agent-lightning.
• Agent Runner: New version providing a more modular and robust runner design.
• Embedded Algorithms: Algorithms are now embedded directly into trainers for simplicity.
• New Execution Strategies: Introduced Client-Server and Shared Memory execution models.
• Trainer Updates: Integrated v0.2 interfaces and FastAlgorithm validation.

Documentation & Examples

• Revamped documentation with new guides for agent creation, training, debugging, and store concepts.
• Improved quickstart tutorials, clarified installation and new deep-dive articles.
• Added and updated examples: SQL Agent, Calc-X, Local SFT, Search-R1, and APO algorithm.

Developer Experience

• Migrated build and CI pipelines to 1ES, split workflows and aggregate badges for clarity.
• Adopted uv as the dependency manager.
• Added GPU-based pytest workflows for full test coverage.
• Enhanced debugging UX, pre-commit configs, and linting (Pyright fixes, import sorting).

Ecosystem & Integrations

• Added support for agents built with Agent-framework.
• Added new community listings: DeepWerewolf and AgentFlow.

New Contributors

A warm welcome to our first-time contributors:
Zhiyuan He (@hzy46), Luna Qiu (@lunaqiu), xjk (@syeehyn), Huanxiang Lin (@linhx1999), Nanako (@SiyunZhao), and Ni Hao (@acured) 🎉

Full changelog: v0.1.2 → v0.2.0

What's Changed

• Add basic documentation in #33
• RAG example by Wang Zilong (@wizardlancet) in #21

New Contributors

• Wang Zilong (@wizardlancet) made their first contribution in #21

Full Changelog: v0.1.1...v0.1.2

What's Changed

• Disable HTTP tracer tests and bump to 0.1.1 in #26
• Fix trainer bugs in v0.1 in #24

Full Changelog: v0.1...v0.1.1

The first release of Agent-lightning!

• Turn your agent into an optimizable beast with ZERO CODE CHANGE (almost)! 💤
• Build with ANY agent framework (LangChain, OpenAI Agent SDK, AutoGen, CrewAI, ...); or even WITHOUT agent framework (Python OpenAI). You name it! 🤖
• Selectively optimize one or more agents in a multi-agent system. 🎯
• Embraces Reinforcement Learning, Automatic Prompt Optimization and more algorithms. 🤗

Install via pip install agentlightning.