Questions About Phoenix vs. LangSmith and Self-Hosted Options

Hi Teams, I have a couple of questions about Phoenix.

· Sorted by Oldest

Xander S.
·
1.
The biggest difference is that Phoenix is open-source and you can self-host it. At the instrumentation level, Phoenix relies on an open standard in the form of OpenInference and OpenTelemetry so you own your telemetry data. We also instrument not only LangChain, but LlamaIndex, OpenAI, and support manual instrumentation via OpenTelemetry for more bespoke use cases.
2.
Yes 🙂
👍1
Sean L.
·
Thank you Xander for confirmation.
🫡1
Jason
·
The Evals library is pretty built out relative to the competitive space, if you are running Evals in code.
👍1
Sean L.
·
I noticed it. Thanks for the hard work. It really helps a lot.
🙌1
Trantor D.
·
I have enabled langsmith alongside Phoenix and it’s seamless really if you want to compare them

Xander S.
·
1.
The biggest difference is that Phoenix is open-source and you can self-host it. At the instrumentation level, Phoenix relies on an open standard in the form of OpenInference and OpenTelemetry so you own your telemetry data. We also instrument not only LangChain, but LlamaIndex, OpenAI, and support manual instrumentation via OpenTelemetry for more bespoke use cases.
2.
Yes 🙂
👍1
Sean L.
·
Thank you Xander for confirmation.
🫡1
Jason
·
The Evals library is pretty built out relative to the competitive space, if you are running Evals in code.
👍1
Sean L.
·
I noticed it. Thanks for the hard work. It really helps a lot.
🙌1
Trantor D.
·
I have enabled langsmith alongside Phoenix and it’s seamless really if you want to compare them