Evaluation, red-team & safety
Eval harnesses, red-teaming, guardrails, and safety tooling for agent systems.
Organic ranking
Free: overall score + blurb. Voice / Builder / Fresh columns unlock with Pro
| # | Tool | Overall |
|---|---|---|
| #1 | LangSmith
Observability, evaluation, and deployment platform for LangChain/LangGraph agents. |
6.80 |
| #2 | Arize Phoenix
Open-source AI observability and eval for traces, embeddings, and LLM apps. |
6.00 |
| #3 | Ragas
Evaluation framework for RAG and LLM applications with reference-free metrics. |
5.80 |