AWS bedrock documentation change

Service: bedrock · 2026-05-07 · Documentation low

File: bedrock/latest/userguide/model-card-amazon-titan-image-generator-g1-v2.md

Summary

Updated sample code to use image generation parameters instead of text inference

Security assessment

Functional change to image generation API without security implications

Diff

diff --git a/bedrock/latest/userguide/model-card-amazon-titan-image-generator-g1-v2.md b/bedrock/latest/userguide/model-card-amazon-titan-image-generator-g1-v2.md
index d151d47fd..0ddc733ca 100644
--- a//bedrock/latest/userguide/model-card-amazon-titan-image-generator-g1-v2.md
+++ b//bedrock/latest/userguide/model-card-amazon-titan-image-generator-g1-v2.md
@@ -95,0 +96 @@ Invoke API
+    import base64
@@ -102,6 +103,8 @@ Invoke API
-                'messages': [{
-                    'role': 'user',
-                    'content': [{'text': 'Can you explain the features of Amazon Bedrock?'}]
-                }],
-                'inferenceConfig': {
-                    'maxTokens': 1024
+                'taskType': 'TEXT_IMAGE',
+                'textToImageParams': {
+                    'text': 'A photograph of a cup of coffee'
+                },
+                'imageGenerationConfig': {
+                    'numberOfImages': 1,
+                    'height': 1024,
+                    'width': 1024
@@ -111 +114,3 @@ Invoke API
-     print(json.loads(response['body'].read()))
+    response_body = json.loads(response['body'].read())
+    base64_image = response_body['images'][0]
+    print(f'Image generated: {len(base64_image)} bytes (base64)')