Error from testing URL Context Tool

Gemini API
,
1

I am testing experimental feature: URL Context tool.

I am experiencing errors that Gemini was unable to access the content, for most of the cases. I’ve tried several news sources.

For example, below is the prompt and output.

Prompt: “Based on the provided url, provide the release time and a summary of the article. url: Russia proposes to hold next talks with Ukraine in Istanbul on June 2 | Reuters”

Output: “I am sorry, but I was unable to access the content of the provided URL to retrieve the release time and summary of the article. This could be due to paywalls, login requirements, or other access restrictions.”

GenerateContentResponse detail:

Candidate content: parts=[Part(video_metadata=None, thought=None, inline_data=None, file_data=None, code_execution_result=None, executable_code=None, function_call=None, function_response=None, text='I am sorry, but I was unable to access the content of the provided URL to retrieve the release time and summary of the article. This could be due to paywalls, login requirements, or other access restrictions.')] role='model'

url_metadata=[UrlMetadata(retrieved_url='https://www.reuters.com/world/europe/russia-proposes-ukraine-date-venue-peace-talks-expects-reply-2025-05-28/', url_retrieval_status=<UrlRetrievalStatus.URL_RETRIEVAL_STATUS_ERROR: 'URL_RETRIEVAL_STATUS_ERROR'>)]

I am looking for more information about which kinds of url are available for using URL Context Tool.

3

Hey @sseo , Welcome to the forum.

The URL you provided seems like it’s behind a paywall or has access restrictions that prevent the tool from retrieving the content. Try using url from government sites, wikipedia or open-access sources that don’t require subscriptions.

Thanks

4

Thanks. I used other AI tools if it can read the url and turned out it also fails. It seems more likely about the policies from the website.

1 Like