Set clipping intervals - doesn't work

Gemini API
, , ,
1

I’m trying to run this example by loading the video first and using video split with intervals (Video understanding  |  Gemini API  |  Google AI for Developers). But it crashes with an error:
Error type: RetryError
Error message: RetryError[<Future at 0x7fe95ea43400 state=finished raised ClientError>]

If I remove the video split, everything is fine. Tested with all 2.5 models. Does it even work for anyone?

3

Hello,

Welcome to the Forum,

Could you please share your code so that I can try to reproduce your issue?

5

Sure!

client = genai.Client(api_key=API_TOKEN)

local_video_path = "vid1.mp4"  # 16sec video, 3.1Mb, 1280 × 720, h264
uploaded_file = client.files.upload(file=local_video_path)

while uploaded_file.state.name == "PROCESSING":
    print(f"File state: {uploaded_file.state.name}. Waiting...")
    time.sleep(2)
    uploaded_file = client.files.get(name=uploaded_file.name)

response = client.models.generate_content(
    model='models/gemini-2.5-pro',
    contents=types.Content(
        parts=[
            types.Part(
                file_data=types.FileData(file_uri=uploaded_file.uri),
                video_metadata=types.VideoMetadata(
                    start_offset='5s',
                    end_offset='15s'
                )
            ),
            types.Part(text='Please summarize the video in 3 sentences.')
        ]
    )
)

print(response.text)

it raised error:

File state: PROCESSING. Waiting...
...
google.genai.errors.ClientError: 400 INVALID_ARGUMENT. {'error': {'code': 400, 'message': 'Unsupported MIME type: ', 'status': 'INVALID_ARGUMENT'}}

The above exception was the direct cause of the following exception:
...
tenacity.RetryError: RetryError[<Future at 0x7f17ed103fd0 state=finished raised ClientError>]

If I run this code without these lines of code:

start_offset='5s',
end_offset='15s'

It works well.

So, I read full logs and found this error: 400 INVALID_ARGUMENT. {'error': {'code': 400, 'message': 'Unsupported MIME type: ', 'status': 'INVALID_ARGUMENT'}}

And I fixed it with adding info mime_type="video/mp4" in types.FileData, so correct code is:

from google import genai
from google.genai import types
import time

API_TOKEN = 

client = genai.Client(api_key=API_TOKEN)

local_video_path = "vid1.mp4"  # 16sec video, 3.1Mb, 1280 × 720, h264
uploaded_file = client.files.upload(file=local_video_path)

start_processing_time = time.time()
while uploaded_file.state.name == "PROCESSING":
    print(f"File state: {uploaded_file.state.name}. Waiting...")
    time.sleep(2)
    uploaded_file = client.files.get(name=uploaded_file.name)

response = client.models.generate_content(
    model='models/gemini-2.5-pro',
    contents=types.Content(
        parts=[
            types.Part(
                file_data=types.FileData(file_uri=uploaded_file.uri, mime_type="video/mp4"),
                video_metadata=types.VideoMetadata(
                    start_offset='5s',
                    end_offset='15s'
                )
            ),
            types.Part(text='Please summarize the video in 3 sentences.')
        ]
    )
)

print(response.text)

Pleas add this information to your documentation here Video understanding  |  Gemini API  |  Google AI for Developers
I also tested it with long 1 hour video, it works well.

P.S. your example with youtube link and offset - works well (without mime_type="video/mp4)

7

Hello,

Apologies for the delayed response and thank you very much for you inputs. I was reproduce the code you provided and observed same.

We will look into the documentation and our code and update the fix as soon as possible.

1 Like