Most-downloaded Observability & Evaluation

26 tools in Observability & Evaluation publish an npm or PyPI package from their own repo. Here they are by downloads over the last 30 days. Langfuse leads with 22.4M.

Counts from the npm downloads API and pypistats; a package only counts if its registry page links back to the tool's repo. Downloads include CI and mirrors, so read them as a usage signal, not a user count. Updated weekly.

#	Tool	Downloads (30 days)	Packages	Per GitHub star
1	Langfuse	22.4M	PyPI `langfuse`	633
2	MLflow	21.4M	PyPI `mlflow` · npm `@mlflow/core`	759
3	Promptfoo	3.0M	npm `promptfoo`	118
4	DeepEval	2.5M	PyPI `deepeval` · npm `deepeval`	137
5	Opik	1.9M	PyPI `opik` · npm `opik`	84
6	MCP Inspector	1.4M	npm `@modelcontextprotocol/inspector`	125
7	Ragas	998.4K	PyPI `ragas`	63
8	phoenix	645.6K	PyPI `arize-phoenix`	55
9	Guardrails	446.5K	PyPI `nemoguardrails`	62
10	OpenLIT	287.2K	PyPI `openlit` · npm `openlit`	102
11	AgentOps	111.1K	PyPI `agentops`	19
12	FastChat	67.2K	PyPI `fschat`	2
13	TensorZero	37.0K	PyPI `tensorzero` · npm `tensorzero`	3
14	garak	36.2K	PyPI `garak`	4
15	agentic-radar	5.3K	PyPI `agentic-radar`	5
16	ChainForge	2.6K	PyPI `chainforge`	1
17	langwatch	1.9K	npm `@langwatch/server`	0
18	UQLM	1.5K	PyPI `uqlm`	1
19	helicone	1.3K	PyPI `helicone` · npm `helicone`	0
20	UpTrain	667	PyPI `uptrain`	0
21	LangFair	505	PyPI `langfair`	2
22	VisionAgent	496	PyPI `vision-agent`	0
23	iFixAi	473	PyPI `ifixai`	0
24	OpenAI Evals	376	PyPI `evals`	0
25	LLM Comparator	47	PyPI `llm-comparator` · npm `llm-comparator`	0
26	Pezzo	16	npm `pezzo`	0

All agent tools by downloads · Observability & Evaluation ranked by GitHub activity