diff --git a/src/autoencoder.test.ts b/src/autoencoder.test.ts
index 166903b6..b649add4 100644
--- a/src/autoencoder.test.ts
+++ b/src/autoencoder.test.ts
@@ -66,3 +66,25 @@ test('test a data sample for anomalies', async () => {
   includesAnomalies(1, 0, 1);
   includesAnomalies(1, 1, 0);
 });
+
+test('restores a net fromJSON', () => {
+  expect(result.error).toBeLessThanOrEqual(errorThresh);
+  function xor(net: AE<number[], number[]>, ...args: number[]) {
+    return Math.round(net.denoise(args)[2]);
+  }
+
+  const json = xornet.toJSON();
+  const net = new AE<number[], number[]>({
+    json,
+  });
+
+  const run1 = xor(net, 0, 0, 0);
+  const run2 = xor(net, 0, 1, 1);
+  const run3 = xor(net, 1, 0, 1);
+  const run4 = xor(net, 1, 1, 0);
+
+  expect(run1).toBe(0);
+  expect(run2).toBe(1);
+  expect(run3).toBe(1);
+  expect(run4).toBe(0);
+});
diff --git a/src/autoencoder.ts b/src/autoencoder.ts
index 357d7fcd..aae06053 100644
--- a/src/autoencoder.ts
+++ b/src/autoencoder.ts
@@ -3,6 +3,7 @@ import {
   IJSONLayer,
   INeuralNetworkData,
   INeuralNetworkDatum,
+  INeuralNetworkJSON,
   INeuralNetworkTrainOptions,
 } from './neural-network';
 import {
@@ -16,6 +17,7 @@ export interface IAEOptions {
   binaryThresh: number;
   decodedSize: number;
   hiddenLayers: number[];
+  json?: INeuralNetworkJSON;
 }
 
 /**
@@ -26,7 +28,7 @@ export class AE<
   EncodedData extends INeuralNetworkData
 > {
   private decoder?: NeuralNetworkGPU<EncodedData, DecodedData>;
-  private readonly denoiser: NeuralNetworkGPU<DecodedData, DecodedData>;
+  private denoiser: NeuralNetworkGPU<DecodedData, DecodedData>;
 
   constructor(options?: Partial<IAEOptions>) {
     // Create default options for the autoencoder.
@@ -47,6 +49,10 @@ export class AE<
 
     // Create the denoiser subnet of the autoencoder.
     this.denoiser = new NeuralNetworkGPU<DecodedData, DecodedData>(options);
+
+    if (options.json) {
+      this.denoiser = this.denoiser.fromJSON(options.json);
+    }
   }
 
   /**
@@ -191,6 +197,15 @@ export class AE<
     return (decoder as unknown) as NeuralNetworkGPU<EncodedData, DecodedData>;
   }
 
+  toJSON(): INeuralNetworkJSON {
+    return this.denoiser.toJSON();
+  }
+
+  fromJSON(json: INeuralNetworkJSON): this {
+    this.denoiser = this.denoiser.fromJSON(json);
+    return this;
+  }
+
   /**
    * Get the layer containing the encoded representation.
    */