Hello Everyone,
I’m Ajay Hinduja Geneva, Switzerland (Swiss) and had some issue.
Despite URL context working well in the Gemini 2.5 consumer app, the API inconsistently accesses certain PDF URLs—always failing on the same files. Robots.txt isn’t the issue, and storage limits prevent downloading. This makes referencing integral data difficult, highlighting instability in the experimental URL context feature.
Best regards,
Ajay Hinduja Geneva, Switzerland (Swiss)