400 'Invalid or unsupported file uri

Gemini API
,
1

Hey, I have been trying to " Provide a list of non function call parts", as per the Python google-genai 1.7.0 documentation, using gemini-2.0-flash, something like:

contents = [
types.Part.from_text(‘What is this image about?’),
types.Part.from_uri(
file_uri: ‘gs://generativeai-downloads/images/scones.jpg’,
mime_type: ‘image/jpeg’,
)
]

but I am getting a 400:

{‘error’: {‘code’: 400, ‘message’: ‘Invalid or unsupported file uri: gs://motorway-condition-centre-vehicle-issues-images/vehicle_images/bd53e001-cc58-43dd-be09-6ecd2ed587a3.jpg’, ‘status’: ‘INVALID_ARGUMENT’}})

I have already checked all obvious potential configs that could cause this issue, like:

  • Incorrect Cloud Storage URI
  • Permissions Issue
  • Incorrect MIME Type
  • VPC Service Controls
  • File Size Limits
  • Invalid characters

What else could be causing this issue?

4

Hi @dedecun , Welcome to the forum.

Just confirming: Are you initializing the client using API key or OAuth using (Vertex AI) ??

5

API Key.

client = genai.Client(api_key=Config.GEMINI_API_KEY)

1 Like
6

You need to initialize the client using OAuth (Vertex AI) in order for the code to work.

client = genai.Client(
    vertexai=True, project='your_project', location='us-central1',
    http_options={'api_version': 'v1'}
)

If you are using the API, you can follow the documentation for vision capabilities.

7

But if I can make this kind of usage work on AI Studio, using API Keys, why not in this case? Also, I used to start it through Vertex, but I was having other issues in my app (like 409 errors with Thinking Models) that were fixed by changing to API Keys. I don’t understand why some features work starting with API Keys and others with Vertex, this makes things quite difficult to understand and work with, not to say very unproductive.