[BUG] Official example not working.

[neoandmatrix]

What happened?

I was trying out the example given at https://github.com/CaviraOSS/OpenMemory/blob/main/docs/getting-started.md for the python but not able to get it work.

Steps to Reproduce

1. create a project with uv init.
2. install openmemory with uv add openmemory-py.
3. paste the given example in main.py
4. run with uv run main.py

Component

Backend (API/Server)

Environment

• OS [Linux/Fedora]
• Python 3.12.9
• Openmemory 1.3.2

Relevant log output

uv run main.py 
Traceback (most recent call last):
  File "/home/iamneo/personal-projects/openmem/main.py", line 3, in <module>
    from openmemory.client import Memory
  File "/home/iamneo/personal-projects/openmem/.venv/lib/python3.12/site-packages/openmemory/__init__.py", line 3, in <module>
    from . import connectors as sources
  File "/home/iamneo/personal-projects/openmem/.venv/lib/python3.12/site-packages/openmemory/connectors/__init__.py", line 5, in <module>
    from .langchain import OpenMemoryChatMessageHistory, OpenMemoryRetriever
  File "/home/iamneo/personal-projects/openmem/.venv/lib/python3.12/site-packages/openmemory/connectors/langchain.py", line 14, in <module>
    class OpenMemoryChatMessageHistory(BaseChatMessageHistory):
  File "/home/iamneo/personal-projects/openmem/.venv/lib/python3.12/site-packages/openmemory/connectors/langchain.py", line 21, in OpenMemoryChatMessageHistory
    def messages(self) -> List[BaseMessage]:
                               ^^^^^^^^^^^
NameError: name 'BaseMessage' is not defined

Code of Conduct

• I agree to follow this project's Code of Conduct

[nullure]

Thanks for reporting the bug, we are working on it!

[NAPTiON]

Docker path issues are a common stumbling block with memory backends. If you want a zero-dependency alternative while this gets fixed: I run a file-based memory pipeline that requires nothing beyond bash + Ollama. No Docker, no SQLite, no embeddings — just a Llama 3.2 1B model categorizing session logs to markdown files on a schedule.

Full setup (scripts included): https://magic.naption.ai/pipeline — might save you time while the Docker issue is being resolved.