# microsoft/agent-lightning

> A lightweight (~3,500 LOC) reinforcement-learning framework for training AI agents with their real harnesses: agents run unchanged while a proxy gateway captures training data. Includes Kubernetes support and coding-agent training examples.

- Magnitude: 1.4 out of 10 — Cooling
- Stars: 18,527 total · +3 stars on 2026-09-29 (UTC), ≈ 7 by evening
- Star trust: star growth looks organic
- Category: AI agents · Language: Python · License: MIT · Created: 2025-06-18 · Last push: 2026-09-29
- GitHub: https://github.com/microsoft/agent-lightning · Homepage: https://microsoft.github.io/agent-lightning/ · Page: https://gitnova.dev/en/r/microsoft/agent-lightning

## Useful for

- Train a coding agent on your own repositories and tests via RL
- Run agent rollouts as Kubernetes Jobs without external sandbox services
- Collect a trajectory dataset from agent model calls via the API Gateway

## Why it’s here

- 3 stars so far today, about 7 expected by the end of the day.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 1,648
- Issues and pull requests: 600
- Watchers: 86
- Average over the last week: 13 per day
- Usual pace: 30 per day
- Latest release: v1.0.2 (2026-09-29)

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-10-04 … 2026-11-02: None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None

## Magnitude by day, last 12 days

09-18 3.1, 09-19 2.5, 09-20 2.5, 09-21 2.2, 09-22 1.9, 09-23 1.6, 09-24 1.4, 09-25 1.4, 09-26 1.4, 09-27 1.6, 09-28 2.1, 09-29 1.4

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-09-29 10:48 UTC, updated every 30 minutes.