diff --git a/blockchain_integration/pi_network/dapi/ai/ml/tensorflow/model.cpp b/blockchain_integration/pi_network/dapi/ai/ml/tensorflow/model.cpp
new file mode 100644
index 000000000..63cab5f33
--- /dev/null
+++ b/blockchain_integration/pi_network/dapi/ai/ml/tensorflow/model.cpp
@@ -0,0 +1,52 @@
+#include "model.h"
+#include <tensorflow/cc/client/client_session.h>
+#include <tensorflow/cc/saved_model/loader.h>
+#include <tensorflow/cc/saved_model/tag_constants.h>
+
+Model::Model(const std::string& model_path) : model_path_(model_path) {}
+
+Model::~Model() {
+  if (session_) {
+    delete session_;
+  }
+}
+
+void Model::load() {
+  // Load the model from a file
+  tensorflow::NewSession(tensorflow::SessionOptions(), &session_);
+  tensorflow::LoadSavedModel(session_, {tensorflow::kSavedModelTagServe}, model_path_, &graph_def_);
+  tensorflow::GraphDef graph_def;
+  TF_CHECK_OK(tensorflow::NewSession(tensorflow::SessionOptions(), &session_));
+  TF_CHECK_OK(session_->Create(graph_def));
+  TF_CHECK_OK(session_->Run({}, {}, {"input"}, &inputs_, &outputs_));
+}
+
+std::vector<float> Model::run(const std::vector<float>& input) {
+  // Create a tensor for the input data
+  tensorflow::Tensor input_tensor(tensorflow::DT_FLOAT, tensorflow::TensorShape({1, input.size()}));
+  auto input_tensor_mapped = input_tensor.tensor<float, 2>();
+  for (int i = 0; i < input.size(); i++) {
+    input_tensor_mapped(0, i) = input[i];
+  }
+
+  // Run the model on the input data
+  std::vector<tensorflow::Tensor> outputs;
+  TF_CHECK_OK(session_->Run({{inputs_[0], input_tensor}}, {"output"}, {}, &outputs));
+
+  // Extract the output data
+  std::vector<float> output;
+  auto output_tensor_mapped = outputs[0].tensor<float, 2>();
+  for (int i = 0; i < output_tensor_mapped.dimension(1); i++) {
+    output.push_back(output_tensor_mapped(0, i));
+  }
+
+  return output;
+}
+
+std::vector<int> Model::get_input_shape() const {
+  return {1, inputs_[0].shape().dim_size(1)};
+}
+
+std::vector<int> Model::get_output_shape() const {
+  return {1, outputs_[0].shape().dim_size(1)};
+}