Langfuse
Open-source LLM engineering platform for agent tracing, prompt management, and evals

Comparing Arize Phoenix (Free OSS (BYOK / self-host), API / SDK) against 10 verified evals, tracing & guardrails alternatives so you can pick the right tool without guessing.
| Tool | Best for | Runs in | Pricing | Setup |
|---|---|---|---|---|
| Arize Phoenixbaseline | Dev teams / BYOK | API / SDK | Free OSS (BYOK / self-host) | CLI / SDK install |
| LangfuseOSS | Dev teams / BYOK | API / SDK | Free 50k obs/mo · Core $29/mo | CLI / SDK install |
| Braintrust | Agent builders | API / SDK | Free tier · Pro $249/mo | CLI / SDK install |
| PromptfooOSS | Dev teams / BYOK | CLI · API / SDK | Free OSS CLI | CLI / SDK install |
| LangSmith | Agent builders | API / SDK | Free tier · Paid plans on site | API / SDK setup |
| HeliconeOSS | Dev teams / BYOK | API / SDK | Free 10k req/mo · Pro $20/seat/mo | API / SDK setup |
| W&B WeaveOSS | Agent builders | API / SDK | Free OSS · Paid cloud tier | CLI / SDK install |
| Comet OpikOSS | Dev teams / BYOK | API / SDK | Free OSS · Paid cloud tier | CLI / SDK install |
| Galileo | Enterprise teams | API / SDK | Custom / enterprise quote | API / SDK setup |
| Lakera Guard | Agent builders | API / SDK | Free tier · Paid plans on site | API / SDK setup |
| Pydantic Logfire | Agent builders | API / SDK | Free tier · Paid plans on site | CLI / SDK install |
Open-source LLM engineering platform for agent tracing, prompt management, and evals
End-to-end evals, trace logging, and PR regression diffs for production AI agents
Open-source CLI for automated LLM evals, agent regression tests, and security red-teaming
Tracing, dataset curation, online evals, and prompt playground for multi-step agents
Open-source LLM proxy and observability platform for cost tracking, caching, and logs
Lightweight tracing, evaluation, and model comparison toolkit from Weights & Biases
Open-source LLM evaluation and high-volume agent trace logging platform
Agent reliability platform for hallucination detection, tool-call scoring, and guardrails
Real-time prompt-injection, jailbreak, and data-leakage firewall for AI agents
OpenTelemetry-native observability platform with first-class PydanticAI agent tracing