@novastera-oss/llamarn 0.1.4-beta.2 → 0.2.1

package/android/CMakeLists.txt

@@ -7,10 +7,10 @@ set(CMAKE_VERBOSE_MAKEFILE ON)
 get_filename_component(MODULE_ROOT "${CMAKE_CURRENT_SOURCE_DIR}/.." ABSOLUTE)
 get_filename_component(CPP_DIR "${MODULE_ROOT}/cpp" ABSOLUTE)
 
-# Define the path to jniLibs - this is where build_android_external.sh puts the libraries
+# Define the path to jniLibs. This assumes CMakeLists.txt is in android/
 set(JNI_LIBS_DIR ${CMAKE_CURRENT_SOURCE_DIR}/src/main/jniLibs)
 
-# Define the path to llama.cpp directory (for headers)
+# Define the path to llama.cpp directory
 set(LLAMA_CPP_DIR "${CPP_DIR}/llama.cpp")
 
 # Make sure the llama.cpp submodule exists
@@ -18,69 +18,40 @@ if(NOT EXISTS "${LLAMA_CPP_DIR}/CMakeLists.txt")
     message(FATAL_ERROR "llama.cpp submodule not found at ${LLAMA_CPP_DIR}. Please run 'git submodule update --init --recursive'")
 endif()
 
-# Check if libraries exist (they should be built by build_android_external.sh)
-if(NOT EXISTS "${JNI_LIBS_DIR}/${ANDROID_ABI}/libllama.so")
-    message(FATAL_ERROR "Prebuilt libraries not found. Please run 'scripts/build_android_external.sh' first to build the native libraries.")
+# Find Vulkan (available in NDK 23+)
+find_package(Vulkan QUIET)
+if(Vulkan_FOUND)
+    message(STATUS "Found Vulkan: ${Vulkan_LIBRARIES}")
+else()
+    message(STATUS "Vulkan not found - GPU acceleration will be limited")
 endif()
 
-# Import all prebuilt libraries as IMPORTED targets
+# Add the prebuilt libraries as IMPORTED with IMPORTED_NO_SONAME to avoid absolute path embedding
 add_library(llama SHARED IMPORTED)
 set_target_properties(llama PROPERTIES 
     IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libllama.so
     IMPORTED_NO_SONAME TRUE)
 
-add_library(ggml SHARED IMPORTED)
-set_target_properties(ggml PROPERTIES 
-    IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml.so
-    IMPORTED_NO_SONAME TRUE)
-
 add_library(ggml-base SHARED IMPORTED)
 set_target_properties(ggml-base PROPERTIES 
     IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-base.so
     IMPORTED_NO_SONAME TRUE)
 
+add_library(ggml SHARED IMPORTED)
+set_target_properties(ggml PROPERTIES 
+    IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml.so
+    IMPORTED_NO_SONAME TRUE)
+
 add_library(ggml-cpu SHARED IMPORTED)
 set_target_properties(ggml-cpu PROPERTIES 
     IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-cpu.so
     IMPORTED_NO_SONAME TRUE)
 
-# Collect additional libraries that exist
-set(ADDITIONAL_LIBRARIES "")
-
-# Check for OpenCL backend
-if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-opencl.so)
-    add_library(ggml-opencl SHARED IMPORTED)
-    set_target_properties(ggml-opencl PROPERTIES 
-        IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-opencl.so
-        IMPORTED_NO_SONAME TRUE)
-    list(APPEND ADDITIONAL_LIBRARIES ggml-opencl)
-    message(STATUS "Found OpenCL backend for ${ANDROID_ABI}")
-endif()
-
-# Check for OpenCL ICD loader
-if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libOpenCL.so)
-    add_library(OpenCL SHARED IMPORTED)
-    set_target_properties(OpenCL PROPERTIES 
-        IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libOpenCL.so
-        IMPORTED_NO_SONAME TRUE)
-    list(APPEND ADDITIONAL_LIBRARIES OpenCL)
-    message(STATUS "Found OpenCL ICD loader for ${ANDROID_ABI}")
-endif()
-
-# Check for Vulkan backend
-if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so)
-    add_library(ggml-vulkan SHARED IMPORTED)
-    set_target_properties(ggml-vulkan PROPERTIES 
-        IMPORTED_LOCATION ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so
-        IMPORTED_NO_SONAME TRUE)
-    list(APPEND ADDITIONAL_LIBRARIES ggml-vulkan)
-    message(STATUS "Found Vulkan backend for ${ANDROID_ABI}")
-endif()
-
-# Create common library with essential llama.cpp common files
+# Create a minimal common library with only essential files that don't require missing GGML symbols
 add_library(
     common
     STATIC
+    # Add back essential files now that we have prebuilt GGML libraries
     ${CPP_DIR}/llama.cpp/common/build-info.cpp
     ${CPP_DIR}/llama.cpp/common/log.cpp
     ${CPP_DIR}/llama.cpp/common/common.cpp
@@ -89,7 +60,6 @@ add_library(
     ${CPP_DIR}/llama.cpp/common/json-schema-to-grammar.cpp
 )
 
-# Create our main React Native module
 add_library(
     RNLlamaCpp
     SHARED
@@ -100,31 +70,61 @@ add_library(
     ${CPP_DIR}/rn-completion.cpp
 )
 
-# Suppress unused function warnings
+# Suppress unused function warnings for llama.cpp code
 target_compile_options(common PRIVATE -Wno-unused-function)
 target_compile_options(RNLlamaCpp PRIVATE -Wno-unused-function)
 
-# Set compile definitions - always enable dynamic loading and CPU backend
+# Check if Vulkan backend library is available
+set(VULKAN_BACKEND_AVAILABLE FALSE)
+if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so)
+    set(VULKAN_BACKEND_AVAILABLE TRUE)
+    message(STATUS "Vulkan backend library found for ${ANDROID_ABI}")
+else()
+    message(STATUS "Vulkan backend library not found for ${ANDROID_ABI}")
+endif()
+
+# Check if OpenCL backend library is available
+set(OPENCL_BACKEND_AVAILABLE FALSE)
+if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-opencl.so AND EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libOpenCL.so)
+    set(OPENCL_BACKEND_AVAILABLE TRUE)
+    message(STATUS "OpenCL backend libraries found for ${ANDROID_ABI}")
+else()
+    message(STATUS "OpenCL backend libraries not found for ${ANDROID_ABI}")
+endif()
+
+# Hybrid backend approach: CPU static (built into main libraries), GPU dynamic
+# CPU backend will be statically linked into main libraries (libggml.so, libllama.so)
+# GPU backends (OpenCL, Vulkan) will be dynamically loaded at runtime only if available
 target_compile_definitions(common PRIVATE 
-    -DGGML_BACKEND_DL=1
-    -DGGML_CPU=1
+    -DGGML_BACKEND_DL=1  # Enable dynamic loading for GPU backends
+    -DGGML_CPU=1         # CPU backend statically built into main libraries
 )
 target_compile_definitions(RNLlamaCpp PRIVATE 
-    -DGGML_BACKEND_DL=1
-    -DGGML_CPU=1
+    -DGGML_BACKEND_DL=1  # Enable dynamic loading for GPU backends
+    -DGGML_CPU=1         # CPU backend statically built into main libraries
 )
 
-# Add GPU backend definitions based on what libraries exist
-if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-opencl.so)
+# DISABLE Vulkan on Android - causes crashes during auto-initialization on emulators
+# Even with n_gpu_layers=0, llama.cpp tries to initialize all available backends
+message(STATUS "Vulkan backend support DISABLED on Android to prevent emulator crashes")
+
+# TODO: Enable Vulkan backend if available (currently disabled due to emulator crashes)
+# Uncomment the lines below to test Vulkan support on real devices
+# if(VULKAN_BACKEND_AVAILABLE)
+#     target_compile_definitions(common PRIVATE -DGGML_VULKAN=1)
+#     target_compile_definitions(RNLlamaCpp PRIVATE -DGGML_VULKAN=1)
+#     message(STATUS "Vulkan backend support enabled for dynamic loading")
+# else()
+#     message(STATUS "Vulkan backend support disabled - library not available")
+# endif()
+
+# Enable OpenCL backend if available
+if(OPENCL_BACKEND_AVAILABLE)
     target_compile_definitions(common PRIVATE -DGGML_OPENCL=1)
     target_compile_definitions(RNLlamaCpp PRIVATE -DGGML_OPENCL=1)
-    message(STATUS "Enabling OpenCL support")
-endif()
-
-if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so)
-    target_compile_definitions(common PRIVATE -DGGML_VULKAN=1)
-    target_compile_definitions(RNLlamaCpp PRIVATE -DGGML_VULKAN=1)
-    message(STATUS "Enabling Vulkan support")
+    message(STATUS "OpenCL backend support enabled for dynamic loading")
+else()
+    message(STATUS "OpenCL backend support disabled - library not available")
 endif()
 
 # Include directories
@@ -133,7 +133,7 @@ target_include_directories(common PRIVATE
     ${LLAMA_CPP_DIR}/ggml/include
     ${LLAMA_CPP_DIR}/include
     ${LLAMA_CPP_DIR}/common
-    ${LLAMA_CPP_DIR}/common/minja
+    ${LLAMA_CPP_DIR}/common/minja  # Add this for chat-template.hpp
     ${LLAMA_CPP_DIR}/src
 )
 
@@ -142,32 +142,44 @@ target_include_directories(RNLlamaCpp PRIVATE
     ${LLAMA_CPP_DIR}/ggml/include
     ${LLAMA_CPP_DIR}/include
     ${LLAMA_CPP_DIR}/common
-    ${LLAMA_CPP_DIR}/common/minja
+    ${LLAMA_CPP_DIR}/common/minja  # Add this for chat-template.hpp
     ${LLAMA_CPP_DIR}/src
     # Add the generated headers path
     ${MODULE_ROOT}/android/generated/jni
     ${MODULE_ROOT}/android/generated/jni/react/renderer/components/RNLlamaCppSpec
 )
 
-# Link libraries - link against all available libraries
+# Link libraries with proper dependencies
 target_link_libraries(
     RNLlamaCpp
     common
-    react_codegen_RNLlamaCppSpec
-    llama
-    ggml-base
-    ggml
-    ggml-cpu
-    ${ADDITIONAL_LIBRARIES}
+    react_codegen_RNLlamaCppSpec  # Link against the generated TurboModule code
+    llama       # Link against the imported prebuilt core llama library
+    ggml-base   # Link against the imported GGML base library
+    ggml        # Link against the imported GGML library
+    ggml-cpu    # Link against the imported GGML CPU library
     jsi
     reactnative
     fbjni
     android
     log
-    dl
+    dl          # Required for dynamic loading of backend libraries
 )
 
-# Copy all libraries to build output directory
+# Add Vulkan support if available
+if(Vulkan_FOUND)
+    target_link_libraries(RNLlamaCpp ${Vulkan_LIBRARIES})
+    target_include_directories(RNLlamaCpp PRIVATE ${Vulkan_INCLUDE_DIRS})
+    message(STATUS "Vulkan support enabled for dynamic GPU backend loading")
+else()
+    # Even without system Vulkan, we can still support dynamic loading if Vulkan library is present at runtime
+    message(STATUS "System Vulkan not found, but dynamic Vulkan loading may still work at runtime")
+endif()
+
+# Add OpenCL support - OpenCL will be loaded dynamically at runtime
+# No need to link against OpenCL here since we use dynamic loading
+
+# Copy dependency libraries to build output directory so they get packaged into APK
 add_custom_command(TARGET RNLlamaCpp POST_BUILD
     COMMAND ${CMAKE_COMMAND} -E copy_if_different
         ${JNI_LIBS_DIR}/${ANDROID_ABI}/libllama.so
@@ -181,16 +193,31 @@ add_custom_command(TARGET RNLlamaCpp POST_BUILD
     COMMAND ${CMAKE_COMMAND} -E copy_if_different
         ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-cpu.so
         $<TARGET_FILE_DIR:RNLlamaCpp>/libggml-cpu.so
-    COMMENT "Copying main libraries"
+    COMMENT "Copying dependency libraries to build output directory"
 )
 
-# Copy any additional GPU libraries that exist
+# Also copy any optional GPU libraries if they exist
+if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so)
+    # Don't copy Vulkan backend on Android - it crashes on emulators during auto-initialization
+    # Even with n_gpu_layers=0, llama.cpp tries to initialize all available backends
+    # and the Android emulator Vulkan driver is broken
+    message(STATUS "Skipping Vulkan backend copy to prevent emulator crashes")
+    
+    # TODO: Uncomment the lines below to enable Vulkan library copying for testing on real devices
+    # add_custom_command(TARGET RNLlamaCpp POST_BUILD
+    #     COMMAND ${CMAKE_COMMAND} -E copy_if_different
+    #         ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so
+    #         $<TARGET_FILE_DIR:RNLlamaCpp>/libggml-vulkan.so
+    #     COMMENT "Copying Vulkan backend library to build output directory"
+    # )
+endif()
+
 if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-opencl.so)
     add_custom_command(TARGET RNLlamaCpp POST_BUILD
         COMMAND ${CMAKE_COMMAND} -E copy_if_different
             ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-opencl.so
             $<TARGET_FILE_DIR:RNLlamaCpp>/libggml-opencl.so
-        COMMENT "Copying OpenCL backend"
+        COMMENT "Copying OpenCL library to build output directory"
     )
 endif()
 
@@ -199,20 +226,11 @@ if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libOpenCL.so)
         COMMAND ${CMAKE_COMMAND} -E copy_if_different
             ${JNI_LIBS_DIR}/${ANDROID_ABI}/libOpenCL.so
             $<TARGET_FILE_DIR:RNLlamaCpp>/libOpenCL.so
-        COMMENT "Copying OpenCL ICD loader"
+        COMMENT "Copying OpenCL loader library to build output directory"
     )
 endif()
 
-if(EXISTS ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so)
-    add_custom_command(TARGET RNLlamaCpp POST_BUILD
-        COMMAND ${CMAKE_COMMAND} -E copy_if_different
-            ${JNI_LIBS_DIR}/${ANDROID_ABI}/libggml-vulkan.so
-            $<TARGET_FILE_DIR:RNLlamaCpp>/libggml-vulkan.so
-        COMMENT "Copying Vulkan backend"
-    )
-endif()
-
-# Expose headers to consuming targets
+# Expose our headers to consuming targets (for autolinking)
 target_include_directories(RNLlamaCpp INTERFACE
     ${CPP_DIR}
     ${LLAMA_CPP_DIR}/ggml/include
@@ -221,4 +239,3 @@ target_include_directories(RNLlamaCpp INTERFACE
     ${LLAMA_CPP_DIR}/common/minja
     ${LLAMA_CPP_DIR}/src
 )
-

package/cpp/LlamaCppModel.cpp

@@ -35,8 +35,8 @@
 
 namespace facebook::react {
 
-LlamaCppModel::LlamaCppModel(rn_llama_context* rn_ctx)
-    : rn_ctx_(rn_ctx), should_stop_completion_(false), is_predicting_(false) {
+LlamaCppModel::LlamaCppModel(rn_llama_context* rn_ctx, std::shared_ptr<CallInvoker> jsInvoker)
+    : rn_ctx_(rn_ctx), should_stop_completion_(false), is_predicting_(false), jsInvoker_(jsInvoker) {
     initHelpers();
 }
 
@@ -435,11 +435,18 @@ CompletionResult LlamaCppModel::completion(const CompletionOptions& options, std
   rn_ctx_->params.n_predict = options.n_predict;
 
   // Check for a partial callback
-  auto callback_adapter = [&partialCallback, runtime](const std::string& token, bool is_done) -> bool {
+  auto callback_adapter = [&partialCallback, runtime, this](const std::string& token, bool is_done) -> bool {
+    // Check for stop condition first
+    if (should_stop_completion_) {
+      return false; // Signal to stop completion
+    }
+    
     if (partialCallback && runtime && !is_done) {
       partialCallback(*runtime, token.c_str());
     }
-    return true;
+    
+    // Return true to continue, false to stop
+    return !should_stop_completion_;
   };
 
   // Run the completion based on whether we have messages or prompt
@@ -546,7 +553,7 @@ jsi::Value LlamaCppModel::jsonToJsi(jsi::Runtime& rt, const json& j) {
   return jsi::Value::undefined();
 }
 
-// JSI method for completions
+// JSI method for completions (synchronous - kept for compatibility)
 jsi::Value LlamaCppModel::completionJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count) {
   if (count < 1 || !args[0].isObject()) {
     throw jsi::JSError(rt, "completion requires an options object");
@@ -581,6 +588,116 @@ jsi::Value LlamaCppModel::completionJsi(jsi::Runtime& rt, const jsi::Value* args
   }
 }
 
+// JSI method for async completions (recommended approach)
+jsi::Value LlamaCppModel::completionAsyncJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count) {
+  if (count < 1 || !args[0].isObject()) {
+    throw jsi::JSError(rt, "completion requires an options object");
+  }
+
+  if (!jsInvoker_) {
+    // Fallback to synchronous if no CallInvoker available
+    return completionJsi(rt, args, count);
+  }
+
+  // Parse options and callback on the current thread
+  CompletionOptions options;
+  std::shared_ptr<jsi::Function> callbackFn = nullptr;
+  
+  try {
+    options = parseCompletionOptions(rt, args[0].getObject(rt));
+    
+    if (count > 1 && args[1].isObject() && args[1].getObject(rt).isFunction(rt)) {
+      callbackFn = std::make_shared<jsi::Function>(args[1].getObject(rt).getFunction(rt));
+      options.stream = true;
+    }
+  } catch (const std::exception& e) {
+    throw jsi::JSError(rt, e.what());
+  }
+
+  // Create Promise constructor
+  auto Promise = rt.global().getPropertyAsFunction(rt, "Promise");
+  
+  auto executor = jsi::Function::createFromHostFunction(
+    rt,
+    jsi::PropNameID::forAscii(rt, "executor"),
+    2,
+    [this, options, callbackFn](jsi::Runtime& runtime, const jsi::Value& thisValue, const jsi::Value* args, size_t count) -> jsi::Value {
+      
+      auto resolve = std::make_shared<jsi::Function>(args[0].asObject(runtime).asFunction(runtime));
+      auto reject = std::make_shared<jsi::Function>(args[1].asObject(runtime).asFunction(runtime));
+      
+      // Create shared references to runtime and invoker for thread safety
+      auto runtimePtr = &runtime;
+      auto invoker = jsInvoker_;
+      auto selfPtr = shared_from_this(); // This requires LlamaCppModel to inherit from std::enable_shared_from_this
+      
+      // Launch background thread for completion
+      std::thread([selfPtr, options, callbackFn, resolve, reject, runtimePtr, invoker]() {
+        try {
+          // Create callback that schedules token updates on JS thread
+          std::function<void(jsi::Runtime&, const char*)> partialCallback = nullptr;
+          
+          if (callbackFn && invoker) {
+            partialCallback = [callbackFn, invoker, runtimePtr](jsi::Runtime& rt, const char* token) {
+              std::string tokenCopy(token);
+              invoker->invokeAsync([callbackFn, tokenCopy, runtimePtr]() {
+                try {
+                  jsi::Object data(*runtimePtr);
+                  data.setProperty(*runtimePtr, "token", jsi::String::createFromUtf8(*runtimePtr, tokenCopy));
+                  callbackFn->call(*runtimePtr, data);
+                } catch (...) {
+                  // Ignore callback errors
+                }
+              });
+            };
+          }
+          
+          // Run completion
+          CompletionResult result = selfPtr->completion(options, partialCallback, runtimePtr);
+          
+          // Schedule success callback on JS thread
+          invoker->invokeAsync([selfPtr, resolve, result, runtimePtr]() {
+            try {
+              jsi::Object jsResult = selfPtr->completionResultToJsi(*runtimePtr, result);
+              resolve->call(*runtimePtr, jsResult);
+            } catch (const std::exception& e) {
+              // If conversion fails, create a simple error response
+              jsi::Object errorObj(*runtimePtr);
+              errorObj.setProperty(*runtimePtr, "error", jsi::String::createFromUtf8(*runtimePtr, e.what()));
+              resolve->call(*runtimePtr, errorObj);
+            }
+          });
+          
+        } catch (const std::exception& e) {
+          // Schedule error callback on JS thread
+          std::string errorMsg(e.what());
+          invoker->invokeAsync([reject, errorMsg, runtimePtr]() {
+            try {
+              reject->call(*runtimePtr, jsi::String::createFromUtf8(*runtimePtr, errorMsg));
+            } catch (...) {
+              // Ignore rejection errors
+            }
+          });
+        }
+      }).detach();
+      
+      return jsi::Value::undefined();
+    }
+  );
+  
+  return Promise.callAsConstructor(rt, std::move(executor));
+}
+
+// JSI method for stopping completion
+jsi::Value LlamaCppModel::stopCompletionJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count) {
+  try {
+    setShouldStopCompletion(true);
+    return jsi::Value(true);
+  } catch (const std::exception& e) {
+    throw jsi::JSError(rt, e.what());
+  }
+}
+
 jsi::Value LlamaCppModel::tokenizeJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count) {
   if (count < 1 || !args[0].isObject()) {
     throw jsi::JSError(rt, "tokenize requires an options object with 'content' field");
@@ -930,12 +1047,36 @@ jsi::Value LlamaCppModel::get(jsi::Runtime& rt, const jsi::PropNameID& name) {
       });
   }
   else if (nameStr == "completion") {
+    // Use async completion as the default to provide better UX
+    if (jsInvoker_) {
+      return jsi::Function::createFromHostFunction(
+        rt, name, 2,
+        [this](jsi::Runtime& runtime, const jsi::Value& thisValue, const jsi::Value* args, size_t count) {
+          return this->completionAsyncJsi(runtime, args, count);
+        });
+    } else {
+      // Fallback to sync completion if no CallInvoker
+      return jsi::Function::createFromHostFunction(
+        rt, name, 2,
+        [this](jsi::Runtime& runtime, const jsi::Value& thisValue, const jsi::Value* args, size_t count) {
+          return this->completionJsi(runtime, args, count);
+        });
+    }
+  }
+  else if (nameStr == "completionSync") {
     return jsi::Function::createFromHostFunction(
       rt, name, 2,
       [this](jsi::Runtime& runtime, const jsi::Value& thisValue, const jsi::Value* args, size_t count) {
         return this->completionJsi(runtime, args, count);
       });
   }
+  else if (nameStr == "stopCompletion") {
+    return jsi::Function::createFromHostFunction(
+      rt, name, 0,
+      [this](jsi::Runtime& runtime, const jsi::Value& thisValue, const jsi::Value* args, size_t count) {
+        return this->stopCompletionJsi(runtime, args, count);
+      });
+  }
   else if (nameStr == "embedding") {
     return jsi::Function::createFromHostFunction(
       rt, name, 1,
@@ -973,6 +1114,8 @@ std::vector<jsi::PropNameID> LlamaCppModel::getPropertyNames(jsi::Runtime& rt) {
   result.push_back(jsi::PropNameID::forAscii(rt, "tokenize"));
   result.push_back(jsi::PropNameID::forAscii(rt, "detokenize"));
   result.push_back(jsi::PropNameID::forAscii(rt, "completion"));
+  result.push_back(jsi::PropNameID::forAscii(rt, "completionSync"));
+  result.push_back(jsi::PropNameID::forAscii(rt, "stopCompletion"));
   result.push_back(jsi::PropNameID::forAscii(rt, "embedding"));
   result.push_back(jsi::PropNameID::forAscii(rt, "release"));
   result.push_back(jsi::PropNameID::forAscii(rt, "n_vocab"));

package/cpp/LlamaCppModel.h

@@ -8,6 +8,9 @@
 #include <unordered_map>
 #include <functional>
 
+// Add ReactCommon includes for proper async handling
+#include <ReactCommon/CallInvoker.h>
+
 // Include all necessary common headers from llama.cpp
 #include "common.h"
 #include "sampling.h"
@@ -70,13 +73,14 @@ struct ToolCall {
  * - Uses common_token_to_piece for token->text conversion
  * - Leverages the llama.cpp chat template system
  */
-class LlamaCppModel : public jsi::HostObject {
+class LlamaCppModel : public jsi::HostObject, public std::enable_shared_from_this<LlamaCppModel> {
 public:
   /**
    * Constructor
    * @param rn_ctx A pointer to an initialized rn_llama_context
+   * @param jsInvoker CallInvoker for async operations (optional, for async completion)
    */
-  LlamaCppModel(rn_llama_context* rn_ctx);
+  LlamaCppModel(rn_llama_context* rn_ctx, std::shared_ptr<CallInvoker> jsInvoker = nullptr);
   virtual ~LlamaCppModel();
 
   /**
@@ -124,6 +128,8 @@ private:
    * JSI method implementations
    */
   jsi::Value completionJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count);
+  jsi::Value completionAsyncJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count);
+  jsi::Value stopCompletionJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count);
   jsi::Value tokenizeJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count);
   jsi::Value detokenizeJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count);
   jsi::Value embeddingJsi(jsi::Runtime& rt, const jsi::Value* args, size_t count);
@@ -157,6 +163,9 @@ private:
   // Completion state
   bool should_stop_completion_;
   bool is_predicting_;
+
+  // Add CallInvoker for async operations
+  std::shared_ptr<CallInvoker> jsInvoker_;
 };
 
 } // namespace facebook::react

package/cpp/PureCppImpl.cpp

@@ -35,7 +35,7 @@ std::shared_ptr<TurboModule> PureCppImpl::create(std::shared_ptr<CallInvoker> js
 }
 
 PureCppImpl::PureCppImpl(std::shared_ptr<CallInvoker> jsInvoker)
-    : NativeRNLlamaCppCxxSpec(std::move(jsInvoker)) {
+    : NativeRNLlamaCppCxxSpec(jsInvoker), jsInvoker_(jsInvoker) {
 }
 
 double PureCppImpl::multiply(jsi::Runtime& rt, double a, double b) {
@@ -325,8 +325,8 @@ jsi::Value PureCppImpl::initLlama(jsi::Runtime &runtime, jsi::Object options) {
 }
 
 jsi::Object PureCppImpl::createModelObject(jsi::Runtime& runtime, rn_llama_context* rn_ctx) {
-  // Create a shared_ptr to a new LlamaCppModel instance
-  auto llamaModel = std::make_shared<LlamaCppModel>(rn_ctx);
+  // Create a shared_ptr to a new LlamaCppModel instance with CallInvoker
+  auto llamaModel = std::make_shared<LlamaCppModel>(rn_ctx, jsInvoker_);
 
   // Create a host object from the LlamaCppModel instance
   return jsi::Object::createFromHostObject(runtime, std::move(llamaModel));

package/cpp/PureCppImpl.h

@@ -54,6 +54,9 @@ private:
 
     // Mutex for thread safety when accessing rn_ctx_ or other shared resources
     std::mutex mutex_;
+    
+    // CallInvoker for async operations
+    std::shared_ptr<CallInvoker> jsInvoker_;
 };
 
 } // namespace facebook::react

package/lib/module/NativeRNLlamaCpp.js.map

@@ -1 +1 @@
-{"version":3,"names":["TurboModuleRegistry","LlamaCppRn","getEnforcing","initLlama","params","loadLlamaModelInfo","modelPath"],"sourceRoot":"../../src","sources":["NativeRNLlamaCpp.ts"],"mappings":";;AACA,SAASA,mBAAmB,QAAQ,cAAc;;AAElD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;;AAiKA;;AA8EA,MAAMC,UAAU,GAAGD,mBAAmB,CAACE,YAAY,CAAO,YAAY,CAAC;AACvE;AACA;AACA;;AAGA;AACA;AACA;AACA,OAAO,SAASC,SAASA,CAACC,MAAwB,EAAmD;EACnG,OAAOH,UAAU,CAACE,SAAS,CAACC,MAAM,CAAC;AACrC;;AAEA;AACA;AACA;AACA,OAAO,SAASC,kBAAkBA,CAChCC,SAAiB,EAWhB;EACD,OAAOL,UAAU,CAACI,kBAAkB,CAACC,SAAS,CAAC;AACjD;AAEA,eAAeL,UAAU","ignoreList":[]}
+{"version":3,"names":["TurboModuleRegistry","LlamaCppRn","getEnforcing","initLlama","params","loadLlamaModelInfo","modelPath"],"sourceRoot":"../../src","sources":["NativeRNLlamaCpp.ts"],"mappings":";;AACA,SAASA,mBAAmB,QAAQ,cAAc;;AAElD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;;AAiKA;;AA+EA,MAAMC,UAAU,GAAGD,mBAAmB,CAACE,YAAY,CAAO,YAAY,CAAC;AACvE;AACA;AACA;;AAGA;AACA;AACA;AACA,OAAO,SAASC,SAASA,CAACC,MAAwB,EAAmD;EACnG,OAAOH,UAAU,CAACE,SAAS,CAACC,MAAM,CAAC;AACrC;;AAEA;AACA;AACA;AACA,OAAO,SAASC,kBAAkBA,CAChCC,SAAiB,EAWhB;EACD,OAAOL,UAAU,CAACI,kBAAkB,CAACC,SAAS,CAAC;AACjD;AAEA,eAAeL,UAAU","ignoreList":[]}

package/lib/typescript/src/NativeRNLlamaCpp.d.ts

@@ -152,6 +152,9 @@ export interface LlamaContextMethods {
     completion(params: LlamaCompletionParams, partialCallback?: (data: {
         token: string;
     }) => void): Promise<LlamaCompletionResult>;
+    completionSync(params: LlamaCompletionParams, partialCallback?: (data: {
+        token: string;
+    }) => void): LlamaCompletionResult;
     tokenize(options: {
         content: string;
         add_special?: boolean;

package/lib/typescript/src/NativeRNLlamaCpp.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"NativeRNLlamaCpp.d.ts","sourceRoot":"","sources":["../../../src/NativeRNLlamaCpp.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAGhD;;;;;;;GAOG;AAEH,MAAM,WAAW,gBAAgB;CAGhC;AAED,MAAM,WAAW,gBAAgB;IAE/B,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,MAAM,CAAC,EAAE,MAAM,CAAC;IAGhB,YAAY,CAAC,EAAE,MAAM,CAAC;IAGtB,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,SAAS,CAAC,EAAE,OAAO,CAAC;IAGpB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,IAAI,CAAC,EAAE,MAAM,CAAC;IAGd,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,eAAe,CAAC,EAAE,MAAM,CAAC;IAGzB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,cAAc,CAAC,EAAE,MAAM,CAAC;IAGxB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,OAAO,CAAC,EAAE,MAAM,CAAC;IAGjB,aAAa,CAAC,EAAE,KAAK,CAAC;QACpB,IAAI,EAAE,MAAM,CAAC;QACb,KAAK,CAAC,EAAE,MAAM,CAAC;KAChB,CAAC,CAAC;IAGH,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,WAAW,qBAAqB;IAEpC,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,EAAE,YAAY,EAAE,CAAC;IAC1B,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC;IAChB,MAAM,CAAC,EAAE,OAAO,CAAC;IAEjB,aAAa,CAAC,EAAE,MAAM,CAAC;IAGvB,WAAW,CAAC,EAAE,MAAM,GAAG,MAAM,GAAG,MAAM,CAAC;IACvC,KAAK,CAAC,EAAE,SAAS,EAAE,CAAC;IAGpB,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,QAAQ,GAAG,MAAM,GAAG,WAAW,GAAG,MAAM,CAAC;IAC/C,OAAO,EAAE,MAAM,CAAC;IAChB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,QAAQ,CAAC;IACf,UAAU,EAAE;QACV,CAAC,GAAG,EAAE,MAAM,GAAG,kBAAkB,CAAC;KACnC,CAAC;IACF,QAAQ,CAAC,EAAE,MAAM,EAAE,CAAC;IACpB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,OAAO,CAAC;IACd,KAAK,EAAE,kBAAkB,CAAC;IAC1B,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,MAAM,gBAAgB,GAAG;IAC7B,IAAI,EAAE,QAAQ,GAAG,QAAQ,GAAG,SAAS,GAAG,MAAM,CAAC;IAC/C,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC;IAChB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB,CAAC;AAEF,MAAM,MAAM,kBAAkB,GAAG,gBAAgB,GAAG,eAAe,GAAG,gBAAgB,CAAC;AAEvF,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,UAAU,CAAC;IACjB,QAAQ,EAAE;QACR,IAAI,EAAE,MAAM,CAAC;QACb,WAAW,CAAC,EAAE,MAAM,CAAC;QACrB,UAAU,EAAE,gBAAgB,CAAC;KAC9B,CAAC;CACH;AAED,MAAM,WAAW,qBAAqB;IACpC,IAAI,EAAE,MAAM,CAAC;IACb,gBAAgB,EAAE,MAAM,CAAC;IACzB,OAAO,EAAE;QACP,WAAW,EAAE,MAAM,CAAC;QACpB,YAAY,EAAE,MAAM,CAAC;QACrB,QAAQ,EAAE,MAAM,CAAC;QACjB,SAAS,EAAE,MAAM,CAAC;QAClB,QAAQ,EAAE,MAAM,CAAC;KAClB,CAAC;IAGF,OAAO,CAAC,EAAE,KAAK,CAAC;QACd,KAAK,EAAE,MAAM,CAAC;QACd,OAAO,EAAE;YACP,IAAI,EAAE,MAAM,CAAC;YACb,OAAO,EAAE,MAAM,CAAC;YAChB,UAAU,CAAC,EAAE,KAAK,CAAC;gBACjB,EAAE,EAAE,MAAM,CAAC;gBACX,IAAI,EAAE,MAAM,CAAC;gBACb,QAAQ,EAAE;oBACR,IAAI,EAAE,MAAM,CAAC;oBACb,SAAS,EAAE,MAAM,CAAC;iBACnB,CAAA;aACF,CAAC,CAAA;SACH,CAAC;QACF,aAAa,EAAE,MAAM,GAAG,QAAQ,GAAG,YAAY,CAAC;KACjD,CAAC,CAAC;IAGH,UAAU,CAAC,EAAE,KAAK,CAAC;QACjB,EAAE,EAAE,MAAM,CAAC;QACX,IAAI,EAAE,MAAM,CAAC;QACb,QAAQ,EAAE;YACR,IAAI,EAAE,MAAM,CAAC;YACb,SAAS,EAAE,MAAM,CAAC;SACnB,CAAC;KACH,CAAC,CAAC;CACJ;AAGD,MAAM,WAAW,gBAAgB;IAC/B,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAAC;IAC1B,OAAO,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAAC;IAC5B,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,eAAe,CAAC,EAAE,OAAO,GAAG,QAAQ,CAAC;IACrC,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,iBAAiB;IAChC,IAAI,EAAE,KAAK,CAAC;QACV,SAAS,EAAE,MAAM,EAAE,GAAG,MAAM,CAAC;QAC7B,KAAK,EAAE,MAAM,CAAC;QACd,MAAM,EAAE,WAAW,CAAC;QACpB,eAAe,CAAC,EAAE,QAAQ,CAAC;KAC5B,CAAC,CAAC;IACH,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE;QACL,aAAa,EAAE,MAAM,CAAC;QACtB,YAAY,EAAE,MAAM,CAAC;KACtB,CAAC;CACH;AAED,MAAM,WAAW,mBAAmB;IAClC,UAAU,CAAC,MAAM,EAAE,qBAAqB,EAAE,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE;QAAC,KAAK,EAAE,MAAM,CAAA;KAAC,KAAK,IAAI,GAAG,OAAO,CAAC,qBAAqB,CAAC,CAAC;IAG7H,QAAQ,CAAC,OAAO,EAAE;QAChB,OAAO,EAAE,MAAM,CAAC;QAChB,WAAW,CAAC,EAAE,OAAO,CAAC;QACtB,WAAW,CAAC,EAAE,OAAO,CAAC;KACvB,GAAG,OAAO,CAAC;QACV,MAAM,EAAE,CAAC,MAAM,GAAG;YAAC,EAAE,EAAE,MAAM,CAAC;YAAC,KAAK,EAAE,MAAM,GAAG,MAAM,EAAE,CAAA;SAAC,CAAC,EAAE,CAAA;KAC5D,CAAC,CAAC;IAGH,UAAU,CAAC,OAAO,EAAE;QAClB,MAAM,EAAE,MAAM,EAAE,CAAA;KACjB,GAAG,OAAO,CAAC;QACV,OAAO,EAAE,MAAM,CAAA;KAChB,CAAC,CAAC;IAEH;;;;;OAKG;IACH,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACjE,cAAc,CAAC,QAAQ,EAAE,YAAY,EAAE,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;IAC1D,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC;IAC5C,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC;IAC5C,cAAc,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;IAChC,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;CAC1B;AAED,MAAM,WAAW,IAAK,SAAQ,WAAW;IAEvC,QAAQ,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAGvC,SAAS,CAAC,MAAM,EAAE,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,GAAG,mBAAmB,CAAC,CAAC;IAGrF,kBAAkB,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAAC;QAC7C,QAAQ,EAAE,MAAM,CAAC;QACjB,OAAO,EAAE,MAAM,CAAC;QAChB,SAAS,EAAE,MAAM,CAAC;QAClB,MAAM,EAAE,MAAM,CAAC;QACf,WAAW,EAAE,MAAM,CAAC;QACpB,YAAY,EAAE,OAAO,CAAC;QACtB,gBAAgB,EAAE,MAAM,CAAC;QACzB,UAAU,CAAC,EAAE,MAAM,CAAC;QACpB,YAAY,CAAC,EAAE,MAAM,CAAC;KACvB,CAAC,CAAC;CACJ;AAED,QAAA,MAAM,UAAU,MAAuD,CAAC;AACxE;;GAEG;AACH,MAAM,MAAM,UAAU,GAAG,gBAAgB,GAAG,mBAAmB,CAAC;AAEhE;;GAEG;AACH,wBAAgB,SAAS,CAAC,MAAM,EAAE,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,GAAG,mBAAmB,CAAC,CAEnG;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAChC,SAAS,EAAE,MAAM,GAChB,OAAO,CAAC;IACT,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,EAAE,MAAM,CAAC;IACf,WAAW,EAAE,MAAM,CAAC;IACpB,YAAY,EAAE,OAAO,CAAC;IACtB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;CACvB,CAAC,CAED;AAED,eAAe,UAAU,CAAC"}
+{"version":3,"file":"NativeRNLlamaCpp.d.ts","sourceRoot":"","sources":["../../../src/NativeRNLlamaCpp.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAGhD;;;;;;;GAOG;AAEH,MAAM,WAAW,gBAAgB;CAGhC;AAED,MAAM,WAAW,gBAAgB;IAE/B,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,MAAM,CAAC,EAAE,MAAM,CAAC;IAGhB,YAAY,CAAC,EAAE,MAAM,CAAC;IAGtB,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,SAAS,CAAC,EAAE,OAAO,CAAC;IAGpB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,IAAI,CAAC,EAAE,MAAM,CAAC;IAGd,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,eAAe,CAAC,EAAE,MAAM,CAAC;IAGzB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,cAAc,CAAC,EAAE,MAAM,CAAC;IAGxB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,OAAO,CAAC,EAAE,MAAM,CAAC;IAGjB,aAAa,CAAC,EAAE,KAAK,CAAC;QACpB,IAAI,EAAE,MAAM,CAAC;QACb,KAAK,CAAC,EAAE,MAAM,CAAC;KAChB,CAAC,CAAC;IAGH,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,WAAW,qBAAqB;IAEpC,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,EAAE,YAAY,EAAE,CAAC;IAC1B,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC;IAChB,MAAM,CAAC,EAAE,OAAO,CAAC;IAEjB,aAAa,CAAC,EAAE,MAAM,CAAC;IAGvB,WAAW,CAAC,EAAE,MAAM,GAAG,MAAM,GAAG,MAAM,CAAC;IACvC,KAAK,CAAC,EAAE,SAAS,EAAE,CAAC;IAGpB,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,QAAQ,GAAG,MAAM,GAAG,WAAW,GAAG,MAAM,CAAC;IAC/C,OAAO,EAAE,MAAM,CAAC;IAChB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,gBAAgB;IAC/B,IAAI,EAAE,QAAQ,CAAC;IACf,UAAU,EAAE;QACV,CAAC,GAAG,EAAE,MAAM,GAAG,kBAAkB,CAAC;KACnC,CAAC;IACF,QAAQ,CAAC,EAAE,MAAM,EAAE,CAAC;IACpB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,OAAO,CAAC;IACd,KAAK,EAAE,kBAAkB,CAAC;IAC1B,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,MAAM,gBAAgB,GAAG;IAC7B,IAAI,EAAE,QAAQ,GAAG,QAAQ,GAAG,SAAS,GAAG,MAAM,CAAC;IAC/C,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC;IAChB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB,CAAC;AAEF,MAAM,MAAM,kBAAkB,GAAG,gBAAgB,GAAG,eAAe,GAAG,gBAAgB,CAAC;AAEvF,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,UAAU,CAAC;IACjB,QAAQ,EAAE;QACR,IAAI,EAAE,MAAM,CAAC;QACb,WAAW,CAAC,EAAE,MAAM,CAAC;QACrB,UAAU,EAAE,gBAAgB,CAAC;KAC9B,CAAC;CACH;AAED,MAAM,WAAW,qBAAqB;IACpC,IAAI,EAAE,MAAM,CAAC;IACb,gBAAgB,EAAE,MAAM,CAAC;IACzB,OAAO,EAAE;QACP,WAAW,EAAE,MAAM,CAAC;QACpB,YAAY,EAAE,MAAM,CAAC;QACrB,QAAQ,EAAE,MAAM,CAAC;QACjB,SAAS,EAAE,MAAM,CAAC;QAClB,QAAQ,EAAE,MAAM,CAAC;KAClB,CAAC;IAGF,OAAO,CAAC,EAAE,KAAK,CAAC;QACd,KAAK,EAAE,MAAM,CAAC;QACd,OAAO,EAAE;YACP,IAAI,EAAE,MAAM,CAAC;YACb,OAAO,EAAE,MAAM,CAAC;YAChB,UAAU,CAAC,EAAE,KAAK,CAAC;gBACjB,EAAE,EAAE,MAAM,CAAC;gBACX,IAAI,EAAE,MAAM,CAAC;gBACb,QAAQ,EAAE;oBACR,IAAI,EAAE,MAAM,CAAC;oBACb,SAAS,EAAE,MAAM,CAAC;iBACnB,CAAA;aACF,CAAC,CAAA;SACH,CAAC;QACF,aAAa,EAAE,MAAM,GAAG,QAAQ,GAAG,YAAY,CAAC;KACjD,CAAC,CAAC;IAGH,UAAU,CAAC,EAAE,KAAK,CAAC;QACjB,EAAE,EAAE,MAAM,CAAC;QACX,IAAI,EAAE,MAAM,CAAC;QACb,QAAQ,EAAE;YACR,IAAI,EAAE,MAAM,CAAC;YACb,SAAS,EAAE,MAAM,CAAC;SACnB,CAAC;KACH,CAAC,CAAC;CACJ;AAGD,MAAM,WAAW,gBAAgB;IAC/B,KAAK,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAAC;IAC1B,OAAO,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAAC;IAC5B,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,eAAe,CAAC,EAAE,OAAO,GAAG,QAAQ,CAAC;IACrC,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,iBAAiB;IAChC,IAAI,EAAE,KAAK,CAAC;QACV,SAAS,EAAE,MAAM,EAAE,GAAG,MAAM,CAAC;QAC7B,KAAK,EAAE,MAAM,CAAC;QACd,MAAM,EAAE,WAAW,CAAC;QACpB,eAAe,CAAC,EAAE,QAAQ,CAAC;KAC5B,CAAC,CAAC;IACH,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE;QACL,aAAa,EAAE,MAAM,CAAC;QACtB,YAAY,EAAE,MAAM,CAAC;KACtB,CAAC;CACH;AAED,MAAM,WAAW,mBAAmB;IAClC,UAAU,CAAC,MAAM,EAAE,qBAAqB,EAAE,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE;QAAC,KAAK,EAAE,MAAM,CAAA;KAAC,KAAK,IAAI,GAAG,OAAO,CAAC,qBAAqB,CAAC,CAAC;IAC7H,cAAc,CAAC,MAAM,EAAE,qBAAqB,EAAE,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE;QAAC,KAAK,EAAE,MAAM,CAAA;KAAC,KAAK,IAAI,GAAG,qBAAqB,CAAC;IAGxH,QAAQ,CAAC,OAAO,EAAE;QAChB,OAAO,EAAE,MAAM,CAAC;QAChB,WAAW,CAAC,EAAE,OAAO,CAAC;QACtB,WAAW,CAAC,EAAE,OAAO,CAAC;KACvB,GAAG,OAAO,CAAC;QACV,MAAM,EAAE,CAAC,MAAM,GAAG;YAAC,EAAE,EAAE,MAAM,CAAC;YAAC,KAAK,EAAE,MAAM,GAAG,MAAM,EAAE,CAAA;SAAC,CAAC,EAAE,CAAA;KAC5D,CAAC,CAAC;IAGH,UAAU,CAAC,OAAO,EAAE;QAClB,MAAM,EAAE,MAAM,EAAE,CAAA;KACjB,GAAG,OAAO,CAAC;QACV,OAAO,EAAE,MAAM,CAAA;KAChB,CAAC,CAAC;IAEH;;;;;OAKG;IACH,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACjE,cAAc,CAAC,QAAQ,EAAE,YAAY,EAAE,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;IAC1D,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC;IAC5C,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC;IAC5C,cAAc,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;IAChC,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;CAC1B;AAED,MAAM,WAAW,IAAK,SAAQ,WAAW;IAEvC,QAAQ,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAGvC,SAAS,CAAC,MAAM,EAAE,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,GAAG,mBAAmB,CAAC,CAAC;IAGrF,kBAAkB,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAAC;QAC7C,QAAQ,EAAE,MAAM,CAAC;QACjB,OAAO,EAAE,MAAM,CAAC;QAChB,SAAS,EAAE,MAAM,CAAC;QAClB,MAAM,EAAE,MAAM,CAAC;QACf,WAAW,EAAE,MAAM,CAAC;QACpB,YAAY,EAAE,OAAO,CAAC;QACtB,gBAAgB,EAAE,MAAM,CAAC;QACzB,UAAU,CAAC,EAAE,MAAM,CAAC;QACpB,YAAY,CAAC,EAAE,MAAM,CAAC;KACvB,CAAC,CAAC;CACJ;AAED,QAAA,MAAM,UAAU,MAAuD,CAAC;AACxE;;GAEG;AACH,MAAM,MAAM,UAAU,GAAG,gBAAgB,GAAG,mBAAmB,CAAC;AAEhE;;GAEG;AACH,wBAAgB,SAAS,CAAC,MAAM,EAAE,gBAAgB,GAAG,OAAO,CAAC,gBAAgB,GAAG,mBAAmB,CAAC,CAEnG;AAED;;GAEG;AACH,wBAAgB,kBAAkB,CAChC,SAAS,EAAE,MAAM,GAChB,OAAO,CAAC;IACT,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;IAChB,SAAS,EAAE,MAAM,CAAC;IAClB,MAAM,EAAE,MAAM,CAAC;IACf,WAAW,EAAE,MAAM,CAAC;IACpB,YAAY,EAAE,OAAO,CAAC;IACtB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;CACvB,CAAC,CAED;AAED,eAAe,UAAU,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@novastera-oss/llamarn",
-  "version": "0.1.4-beta.2",
+  "version": "0.2.1",
   "description": "An attempt at a pure cpp turbo module library",
   "source": "./src/index.tsx",
   "main": "./lib/module/index.js",
@@ -66,6 +66,7 @@
     "build-android-external": "bash scripts/build_android_external.sh",
     "build-android-clean": "bash scripts/build_android_external.sh --clean-prebuilt",
     "build-android-macos": "npm run clean-android && npm run clean-prebuilt && bash scripts/build_android_gpu_backend.sh --no-vulkan --no-opencl --clean && bash scripts/build_android_external.sh --no-vulkan --no-opencl --clean"
+  
   },
   "keywords": [
     "react-native",

package/src/NativeRNLlamaCpp.ts

@@ -195,6 +195,7 @@ export interface EmbeddingResponse {
 
 export interface LlamaContextMethods {
   completion(params: LlamaCompletionParams, partialCallback?: (data: {token: string}) => void): Promise<LlamaCompletionResult>;
+  completionSync(params: LlamaCompletionParams, partialCallback?: (data: {token: string}) => void): LlamaCompletionResult;
 
   // Updated tokenize method to match server.cpp interface
   tokenize(options: {