Unfortunately, yes. Not by choice.

What are some alternatives? I've been tinkering with langgraph as of late and frankly the whole space is so polluted with SEO and vibe-coded systems the old "classics" were the safe bet for me.

Langchain the company makes three different main products, all of which are differing levels of bad in my experience.

LangGraph, for agent/workflow orchestration is the least bad of the three, but has solid alternatives these days, such as OpenAI’s own Agents SDK, or Pydantic AI.

LangSmith, the platform for prompt authoring/experimentation/observability isn’t great but is useable. I would much prefer Langfuse over it at this point.

Langchain, the library for interfacing with LLMs, is absolutely terrible. There is virtually nothing good to say about it, and in fact is so bad that the fact that LangSmith more or less requires you to use Langchain to some degree is probably the biggest knock against it. Langchain doesn’t even need an alternative, literally just interfacing with the LLMs directly through their clients are often simpler, more flexible, and preferable to using Langchain.

Is langfuse the best? It seems nice but I'm not sure what the options are.

For other open-source alternatives, check out https://github.com/agenta-ai/agenta (I am a maintainer), we're more focused on evals and collaboration between subject matter experts and engineers (friendlier UI, powerful prompt engineering flow (playground, eval from UI, etc..)).

There are so many products in the space right now, I hesitate to say “best” because I have only scratched the surface. It’s the best that I’ve tried. The main things it has going for it:

- It’s open source

- It has the right level of abstraction IMO. A lot of products in the space are over engineered, which make them brittle and a pain to work with. Langfuse is focused enough that it does what it needs to well, but is flexible enough to use with a lot of other tools.

- The developers seem competent. Their roadmap looks solid and they develop very quickly for a team their size.

- It works. I can’t tell you the number of times I’ve tried to do something with Langchain products where it just breaks doing simple things. Lots of bugs, bad documentation. Langfuse is better in this regard.

- It’s cheaper (at least it is cheaper than LangSmith). LangSmith charges you for seats and per trace. They charge you even more if you want to keep a trace for more than 14 days, and even then you can only keep traces for a max of 400 days, even if you are hosting on your own servers. Langfuse charges you for seats and that’s it. If you are self-hosting you can do what you want with your data.

There are possibly other platforms that are similar to Langfuse as well, but it’s the best I’ve encountered so far.

basically you build your own by picking and choosing the parts that makes sense for your use case (often with ai)

and you end up with xkcd 927 (standards)

>Are people still building with lang chain?

Prime example of the first mover advantage. It's a complete dumpster fire, but they caught enough mindshare early on to be thought of as the default choice now for people newly coming into the agentic world.