10 GitHub repos to see what your AI agent really did, from Langfuse to claude-code-log
victor_explore · x · 2026-10-16
A curated list of 10 GitHub projects for tracing and observing what AI agents actually do (stars as of Oct 2026):
- Langfuse (36K stars, MIT core): traces every LLM call and tool step with cost, time and scores; self-hostable.
- MLflow (28K stars, Apache 2.0): the classic ML platform, now tracing LLM apps and agents with evals.
- Opik (22K stars, Apache 2.0): Comet's tool to trace, score and monitor LLM apps.
- Phoenix (12K stars, ELv2): Arize's OpenTelemetry-based tracing and eval app, free to self-host but not open source.
- OpenLLMetry (7K stars): adds OpenTelemetry tracing so traces flow to any existing tool.
- LangWatch (5K stars): traces, evals and agent simulations.
- Also: Laminar (3K stars, agent-focused tracing), OpenLIT (3K stars, one-line OpenTelemetry), Claude Code History Viewer (2K stars), and claude-code-log (1K stars, converts Claude Code transcripts to readable HTML).
More from coding & agent
- MLExpert: 35B MoE Agent Trained to Autonomously Handle Machine Learning Engineering — Shangheng Du · 2026-10-16
- Cowork-Gym: Alibaba Synthesizes 64.9K Agent Work Environments, Boosting Qwen by 8.7 Points — Anni Zou · 2026-10-16
- MMSearch-OPD: Joint RL and On-Policy Distillation for Long-Horizon Multimodal Search Agents — PekingUniversity · 2026-10-16
- A gallery of 214 projects built on one model shows the 'post your build' ad flywheel — gaganghotra_ · 2026-10-16
- OpenAI discloses GPT-5.6 Sol derived wallet keys and spent unauthorized funds — terryyuezhuo · 2026-10-16
- WSL 3.0 ships with native containers and Linux 6.18 kernel, syscall overhead down ~16% — solyarisoftware · 2026-10-16