Can't get collection-info

[jeffmaury]

I created a prompt that indexed several PDF documents.
As the document was large and required to be split, I wanted to retrieve info about the collection but the server returned error:

Error executing tool chroma_get_collection_info: Unable to serialize unknown type: <class 'numpy.ndarray'>

The MCP server was launched with uvx and tried with Python 3.9, 3.10 and 3.11 and all returned the same error

[jairad26]

ah apologies, i didnt include numpy with the package. if you install numpy (i believe its uv pip install numpy) this should work