Deployment Error: OpenAIEmbedding Attribute Error

LangSmith Product Help Deployment
1

I am writing to you regarding a deployment issue we are facing on the LangSmith Platform.

When attempting a deployment on the LangSmith platform, we have been encountering an Attribute Error related to OpenAIEmbedding, which is preventing the deployment from completing.
The details of the development environment where the issue is occurring are as follows:

  • Development ID: 7af5187a-6cdb-4d1b-b6b9-250eaf6f72ac
  • Representative Revision with the Issue: a31470e6-5b72-4100-9004-c29e7b401109

logs:

2025/12/23 11:30:13
[INFO] AttributeError: 'OpenAIEmbeddings' object has no attribute 'strip'
2025/12/23 11:30:13
[INFO]     raise AttributeError(f'{type(self).__name__!r} object has no attribute {item!r}')
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/pydantic/main.py", line 1026, in __getattr__
2025/12/23 11:30:13
[INFO]            ^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     if not model.strip():
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/langchain/embeddings/base.py", line 86, in _infer_model_and_provider
2025/12/23 11:30:13
[INFO]                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     provider, model_name = _infer_model_and_provider(model, provider=provider)
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/langchain/embeddings/base.py", line 188, in init_embeddings
2025/12/23 11:30:13
[INFO]   File "/api/langgraph_api/graph.py", line 708, in resolve_embeddings
2025/12/23 11:30:13
[INFO]   File "/api/langgraph_api/timing/timer.py", line 101, in wrapper
2025/12/23 11:30:13
[INFO]                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     _STORE_CONFIG["index"]["embed"] = resolve_embeddings(_STORE_CONFIG["index"])
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/langgraph_runtime_postgres/store.py", line 130, in set_store_config
2025/12/23 11:30:13
[INFO]     lg_store.set_store_config(config_)
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/langgraph_runtime_postgres/database.py", line 209, in migrate_vector_index
2025/12/23 11:30:13
[INFO]     await migrate_vector_index()
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/langgraph_runtime_postgres/database.py", line 237, in start_pool
2025/12/23 11:30:13
[INFO]     await database.start_pool()
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/site-packages/langgraph_runtime_postgres/lifespan.py", line 68, in lifespan
2025/12/23 11:30:13
[INFO]            ^^^^^^^^^^^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     return await anext(self.gen)
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/contextlib.py", line 210, in __aenter__
2025/12/23 11:30:13
[INFO]              ^^^^^^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     result = await _enter(cm)
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/contextlib.py", line 650, in enter_async_context
2025/12/23 11:30:13
[INFO]   File "/api/langgraph_api/timing/timer.py", line 227, in combined_lifespan
2025/12/23 11:30:13
[INFO]            ^^^^^^^^^^^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     return await anext(self.gen)
2025/12/23 11:30:13
[INFO]   File "/usr/lib/python3.11/contextlib.py", line 210, in __aenter__
2025/12/23 11:30:13
[INFO]              ^^^^^^^^^^^^^^^^
2025/12/23 11:30:13
[INFO]     result = await _enter(cm)
2025/12/23 11:30:13
[INFO] Traceback (most recent call last):
  File "/usr/lib/python3.11/site-packages/langgraph_runtime_postgres/lifespan.py", line 168, in lifespan
    yield
  File "/api/langgraph_api/timing/timer.py", line 227, in combined_lifespan
  File "/usr/lib/python3.11/contextlib.py", line 650, in enter_async_context

For the langgraph.json file, we are currently using the following settings:

"store": {
  "index": {
    "embed": "openai:text-embedding-3-small",
    "dims": 1536,
    "fields": ["content"]
  }
},

For your reference, I will also add the response from Codex based on the logs:

Root Cause Analysis from Internal Investigation (Response from Codex):
The definitive cause is that an OpenAIEmbeddings object, rather than a string (e.g., “openai:text-embedding-3-small”), is present in _STORE_CONFIG["index"]["embed"] at the time langgraph_runtime_postgres.store.set_store_config() is executed. This leads to an AttributeError when model.strip() is called inside langchain.embeddings.base._infer_model_and_provider() , as the strip() method does not exist for the OpenAIEmbeddings object. This issue is unrelated to the application’s graph execution and occurs during the Platform’s startup Store initialization.

2

Thanks for flagging - will look into this

1 Like
3

@Ren , do you still have issues if you re-deploy? Agent server version 0.6.12 should address the issue you’re seeing.

5

@wfh

Thanks a lot for the fast fix!

I tried redeploying and it worked perfectly.
The issue is all cleared up. I’m impressed that you handled this in only a few hours.

Did you guys update the langgraph-api version to 0.6.12 to fix this?

Regardless, thanks a million for the help!

6

Yes exactly !

1 Like