Protect your Firebase ML Android app's Cloud credentials

If your Android app uses one of Firebase ML's cloud APIs, before you launch your app in production, you should take some additional steps to prevent unauthorized API access.

For your production apps, you will ensure that only authenticated clients can access cloud services. (Note that only non-rooted devices can authenticate using the method described.)

Then, you will create a debug-only API key that you can use for convenience during testing and development.

1. Register your production apps with Firebase

First, register your production apps with Firebase.

Make sure that you have your app's SHA-1 signatures. Refer to Authenticating your client to learn how.
Go to your Project settings in the Firebase console, then select the Settings tab.
Scroll down to the Your apps card, then select your Android app.
Add your app's SHA-1 signature to your app's information.

2. Restrict the scope of your API keys

Next, configure your existing API keys to disallow access to the Cloud Vision API:

Open the Credentials page of the Google Cloud Console. When prompted, select your project.
For each existing API key in the list, open the editing view.
In the API restrictions section, select Restrict key, then add to the list all of the APIs to which you want the API key to have access. Make sure to not include the Cloud Vision API.

When you configure an API key's API restrictions, you are explicitly declaring the APIs to which the key has access. By default, when the API restrictions section has Don't restrict key selected, an API key can be used to access any API that is enabled for the project.

Now, your existing API keys will not grant access to cloud ML services, but each key will continue to work for any APIs that you added to its API restrictions list.

Note that if you enable any additional APIs in the future, you must add them to the API restrictions list for the applicable API key.

3. Create and use a debug-only API key

Finally, create a new API key to be used only for development. Firebase ML can use this API key to access Google Cloud services in environments where app authentication isn't possible, such as when running on emulators.

Create a new API key to be used for development:
1. Open the Credentials page of the Google Cloud Console. When prompted, select your project.
2. Click Create credentials > API key and take note of the new API key. This key allows API access from unauthenticated apps, so keep this key confidential.
To ensure the new debug API key is not leaked with your released app, specify the debug API key in an Android manifest file used only for debug builds:
1. If you don't already have a debug manifest, create one by clicking File > New > Other > Android Manifest File and selecting debug from the target source sets.
2. In the debug manifest, add the following declaration:
```
<application>
<meta-data
    android:name="com.firebase.ml.cloud.ApiKeyForDebug"
    android:value="your-debug-api-key" />
</application>
```

In your app, configure Firebase ML to use certificate fingerprint matching to authenticate your client in production and to use API keys—the debug key—only in debug builds:

Kotlin+KTX

val optionsBuilder = FirebaseVisionCloudImageLabelerOptions.Builder()
if (!BuildConfig.DEBUG) {
    // Requires physical, non-rooted device:
    optionsBuilder.enforceCertFingerprintMatch()
}

// Set other options. For example:
optionsBuilder.setConfidenceThreshold(0.8f)
// ...

// And lastly:
val options = optionsBuilder.build()
FirebaseVision.getInstance().getCloudImageLabeler(options).processImage(myImage)MainActivity.kt

Java

FirebaseVisionCloudImageLabelerOptions.Builder optionsBuilder =
        new FirebaseVisionCloudImageLabelerOptions.Builder();
if (!BuildConfig.DEBUG) {
    // Requires physical, non-rooted device:
    optionsBuilder.enforceCertFingerprintMatch();
}

// Set other options. For example:
optionsBuilder.setConfidenceThreshold(0.8f);
// ...

// And lastly:
FirebaseVisionCloudImageLabelerOptions options = optionsBuilder.build();
FirebaseVision.getInstance().getCloudImageLabeler(options).processImage(myImage);MainActivity.java

Next steps

See the launch checklist for information on preparing your app to launch when using other Firebase features.