I’m looking for guidance on best practices for building an image to run LangGraph Server standalone using the OSS version in my own infrastructure.
It seems that using the LangGraph CLI image still requires LangGraph keys, even when deploying on my own infrastructure. Could you clarify the recommended approach for self-hosted deployments, and whether there is a way to run it without relying on LangGraph keys?
Thanks
Hello @mag,
Unfortunately, there is currently no way to use a self-hosted deployment unless you are an enterprise customer with access to the LangGraph keys.
I already provided a comprehensive answer to another community member about this query here, so I’d recommend checking it out, as this is a very common point of confusion.
I saw that, but I didn’t find any guidance on deploying the OSS server without the keys. Are there any best practices for implementing a layer that avoids the components requiring a license key?
We chose LangGraph because it is advertised as open source framework, so any clarification on how to run it in a fully self-hosted setup would be appreciated.
Thanks
Hello @mag, Unfortunately, right now, there is no best way that I am aware of to implement the agent server in production without bypassing the license key. You can still start a local dev server which replicates full production-grade readiness using langgraph up, but that is not recommended in production.
Note: Langraph OSS itself is free to use, but the associated binaries that run the Langgraph server are an enterprise solution which requires an enterprise key to run in production.
For further clarification, I would recommend you delve into the following documentation so this confusion get’s cleared up: Set up LangSmith - Docs by LangChain so you clearly understand the different deployment options you have for Langgraph servers.
@mag let me know if this is clear or if there is still some confusion?