Hey, I’m trying to get the gemini api to fetch a url which is a pinggy link (it’s like ngrok) that goes to a local server i have that just returns a hello world html page - this is just to test urlContext capabilities. When I try the query in gemini.google.com it works perfectly fine, i see the fetch happen on my server, and the it outputs the html it got back. But locally I don’t even see the request come in to my server - this is the API request I am making:
{
model: 'models/gemini-3-pro-preview',
contents: [
{
role: 'user',
parts: [
{
text: 'Get this url pls https://rzgfgwnpjl.a.pinggy.link/op/get/APP01 - it is tunnel so you can access it. Summarize its content for me.'
}
]
}
],
config: {
maxOutputTokens: 54000,
stopSequences: undefined,
abortSignal: AbortSignal { aborted: false },
temperature: 0.7,
thinkingConfig: { includeThoughts: true, thinkingLevel: 'HIGH' },
systemInstruction: 'You are a helpful assistant',
tools: [ { googleSearch: {} }, { urlContext: {} } ]
}
}
I have tried asking gemini further messages for that it tried and what it saw, and it looks like:
**Tool Used:** `google_search` (specifically the `browse` function for URL extraction)
**Parameters:** `urls=['https://rzgfgwnpjl.a.pinggy.link/op/get/APP01']`
**Result:** `Not able to browse the provided URL. This maybe because the provided URL is blocked.`
Another attempt telling it to use the browse tools game me:
* **Tool Used:** `google_browse`
* **Parameters Used:** `{'urls': ['https://rzgfgwnpjl.a.pinggy.link/op/get/APP01']}`
* **Verbatim Result:** "Not able to browse the provided URL. This maybe because the provided URL is blocked."
Any ideas here? Will this simply just not work with urlContext?