I am developing an Android application in Google AI Studio and have noticed unusually high network/data consumption during relatively small code-editing tasks.
In one recent case, a relatively small PDF viewer modification appeared to consume approximately 400 MB or more of network data, even though:
- the application itself is small;
- the source project is relatively small;
- the requested change was limited to a single source file;
- the change did not involve downloading a large external asset or dependency.
This makes the current workflow difficult for users on limited or metered connections.
I would like Google to investigate whether AI Studio is repeatedly transferring project state, build artifacts, source snapshots, previews, generated outputs, or other resources during these operations.
Important product suggestion
I strongly suggest making Preview and Code resource usage more explicitly demand-driven:
-
When the user switches to Code and is only inspecting code or downloading the APK, Preview-related resources should not continue downloading or processing in the background unless Preview is actively requested.
-
When the user leaves Preview and switches to Code, Preview-specific background loading/processing should pause or stop and resume only when the user returns to Preview.
-
Restore/snapshot data should not be downloaded, reconstructed, or processed unless the user explicitly initiates a Restore operation.
-
Build/download operations requested from Code should avoid unnecessary Preview-related data transfer.
-
The interface would benefit from clearer indication of what background operations are currently consuming network bandwidth.
Why this matters
Repeated background transfers are especially problematic for users with limited mobile data plans.
Please investigate the network/data usage associated with:
- Preview;
- Code;
- APK builds/downloads;
- project snapshots/history;
- Restore operations;
- background synchronization.
I would appreciate clarification on what is actually being transferred during these actions and whether AI Studio can reduce unnecessary background traffic.