expo-ai-kit 0.3.1 → 0.3.3

package/README.md

@@ -2,7 +2,7 @@
 
 On-device AI for Expo apps. Run language models locally—no API keys, no cloud, just native intelligence.
 
-**Now with Gemma 4 support** — Download and run Google's [Gemma 4](https://blog.google/technology/developers/gemma-4/) E2B (2.3B) and E4B (4.5B) models directly on Android devices via [LiteRT-LM](https://ai.google.dev/edge/litert-lm). Full on-device inference with GPU acceleration, streaming, and zero cloud dependency.
+**Now with Gemma 4 support** — Download and run Google's [Gemma 4](https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/) E2B (2.3B) and E4B (4.5B) models directly on Android devices via [LiteRT-LM](https://ai.google.dev/edge/litert-lm). Full on-device inference with GPU acceleration, streaming, and zero cloud dependency.
 
 [![npm version](https://img.shields.io/npm/v/expo-ai-kit.svg)](https://www.npmjs.com/package/expo-ai-kit)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)

package/android/build.gradle

@@ -44,5 +44,5 @@ android {
 
 dependencies {
   implementation "com.google.mlkit:genai-prompt:1.0.0-beta2"
-  implementation "com.google.ai.edge.litertlm:litertlm-android:+"
+  implementation "com.google.ai.edge.litertlm:litertlm-android:0.8.0"
 }

package/android/src/main/java/expo/modules/aikit/ExpoAiKitModule.kt

@@ -208,7 +208,7 @@ class ExpoAiKitModule : Module() {
       activeModelId
     }
 
-    AsyncFunction("unloadModel") Coroutine {
+    AsyncFunction("unloadModel") Coroutine { ->
       if (activeModelId != "mlkit" && gemmaClient.isModelLoaded()) {
         val previousId = activeModelId
         gemmaClient.unloadModel()

package/android/src/main/java/expo/modules/aikit/GemmaInferenceClient.kt

@@ -5,6 +5,7 @@ import com.google.ai.edge.litertlm.Engine
 import com.google.ai.edge.litertlm.EngineConfig
 import com.google.ai.edge.litertlm.Conversation
 import com.google.ai.edge.litertlm.Backend
+import com.google.ai.edge.litertlm.Message
 import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.sync.Mutex
 import kotlinx.coroutines.sync.withLock
@@ -64,7 +65,7 @@ class GemmaInferenceClient(private val context: Context) {
       withContext(Dispatchers.IO) {
         val engineConfig = EngineConfig(
           modelPath = modelPath,
-          backend = Backend.GPU()
+          backend = Backend.GPU
         )
         val newEngine = Engine(engineConfig)
         newEngine.initialize()
@@ -122,7 +123,7 @@ class GemmaInferenceClient(private val context: Context) {
 
     try {
       withContext(Dispatchers.IO) {
-        conv.sendMessage(contents = fullPrompt).toString()
+        conv.sendMessage(Message.of(fullPrompt)).toString()
       }
     } catch (e: OutOfMemoryError) {
       throw RuntimeException("INFERENCE_OOM:${loadedModelId ?: "unknown"}:Out of memory during inference")
@@ -153,7 +154,7 @@ class GemmaInferenceClient(private val context: Context) {
       withContext(Dispatchers.IO) {
         var previousText = ""
 
-        conv.sendMessageAsync(contents = fullPrompt).collect { message ->
+        conv.sendMessageAsync(Message.of(fullPrompt)).collect { message ->
           val accumulated = message.toString()
           val token = if (accumulated.length > previousText.length) {
             accumulated.substring(previousText.length)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "expo-ai-kit",
-  "version": "0.3.1",
+  "version": "0.3.3",
   "description": "On-device AI for Expo apps — run Gemma 4, Apple Foundation Models, and ML Kit locally with zero API keys",
   "main": "build/index.js",
   "types": "build/index.d.ts",