feat(ai): Redact base64 data URLs in image_url content blocks

[ericapisani]

Extend redact_blob_message_parts to detect and redact base64 data URLs inside image_url content blocks (e.g. data:image/jpeg;base64,...), in addition to the existing blob type handling.

Some AI integrations send image content as image_url items with inline base64 data URLs rather than the blob content type. Without this change, those base64 payloads are sent as span data, which inflates event size and can leak image content.

Also moves DATA_URL_BASE64_REGEX from sentry_sdk/integrations/pydantic_ai/consts.py to sentry_sdk/ai/consts.py since it's now shared across AI monitoring code beyond pydantic_ai.

Fixes PY-2280 and #5948

[linear-code]

PY-2280 Ensure that `gen_ai.request.messages` don't contain base64 strings

[github-actions]

Semver Impact of This PR

🟡 Minor (new features)

📋 Changelog Preview

This is how your changes will appear in the changelog.
Entries from this PR are highlighted with a left border (blockquote style).

---

New Features ✨

• (ai) Redact base64 data URLs in image_url content blocks by ericapisani in #5953

• (integrations) Instrument pyreqwest tracing by servusdei2018 in #5682

Internal Changes 🔧

• (openai) Split token counting by API for easier deprecation by ericapisani in #5930
• (opentelemetry) Ignore mypy error by alexander-alderman-webb in #5927
• Fix license metadata in setup.py by sl0thentr0py in #5934
• Update validate-pr workflow by stephanie-anderson in #5931

---

_{🤖 This preview updates automatically when you update the PR.}

[ericapisani]

bugbot run

[github-actions]

Codecov Results 📊

✅ 13 passed | Total: 13 | Pass Rate: 100% | Execution Time: 9.77s

All tests are passing successfully.

❌ Patch coverage is 25.00%. Project has 14842 uncovered lines.

Files with missing lines (3)

File	Patch %	Lines
utils.py	15.25%	⚠️ 239 Missing
utils.py	0.00%	⚠️ 27 Missing
consts.py	0.00%	⚠️ 1 Missing

---

Generated by Codecov Action

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Autofix prepared a fix for the issue found in the latest run.

• ✅ Fixed: Crashes when image_url value is a string shorthand
  • Added type checking to handle both string shorthand and dict formats for image_url in _is_image_type_with_blob_content and redact_blob_message_parts functions, preventing AttributeError when calling .get() on a string.

Or push these changes by commenting:

@cursor push cde852a515

Preview (cde852a515)

diff --git a/sentry_sdk/ai/utils.py b/sentry_sdk/ai/utils.py
--- a/sentry_sdk/ai/utils.py
+++ b/sentry_sdk/ai/utils.py
@@ -597,7 +597,14 @@
     if item.get("type") != "image_url":
         return False
 
-    image_url = item.get("image_url", {}).get("url", "")
+    image_url_data = item.get("image_url")
+    if isinstance(image_url_data, str):
+        image_url = image_url_data
+    elif isinstance(image_url_data, dict):
+        image_url = image_url_data.get("url", "")
+    else:
+        return False
+
     data_url_match = DATA_URL_BASE64_REGEX.match(image_url)
 
     return bool(data_url_match)
@@ -682,7 +689,11 @@
                     if item.get("type") == "blob":
                         item["content"] = BLOB_DATA_SUBSTITUTE
                     elif _is_image_type_with_blob_content(item):
-                        item["image_url"]["url"] = BLOB_DATA_SUBSTITUTE
+                        image_url_data = item.get("image_url")
+                        if isinstance(image_url_data, str):
+                            item["image_url"] = BLOB_DATA_SUBSTITUTE
+                        elif isinstance(image_url_data, dict):
+                            item["image_url"]["url"] = BLOB_DATA_SUBSTITUTE
 
     return messages_copy
 

diff --git a/tests/test_ai_monitoring.py b/tests/test_ai_monitoring.py
--- a/tests/test_ai_monitoring.py
+++ b/tests/test_ai_monitoring.py
@@ -845,6 +845,38 @@
         assert result[0]["content"][1]["type"] == "image_url"
         assert result[0]["content"][1]["image_url"]["url"] == BLOB_DATA_SUBSTITUTE
 
+    def test_redacts_image_url_string_shorthand_with_blob(self):
+        """Test redacting image_url using string shorthand format with base64 data"""
+        messages = [
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "text": "How many ponies do you see in the image?",
+                        "type": "text",
+                    },
+                    {
+                        "type": "image_url",
+                        "image_url": "data:image/jpeg;base64,/9j/4AAQSkZJRg==",
+                    },
+                ],
+            }
+        ]
+
+        original_blob_content = messages[0]["content"][1]
+
+        result = redact_blob_message_parts(messages)
+
+        assert messages[0]["content"][1] == original_blob_content
+
+        assert (
+            result[0]["content"][0]["text"]
+            == "How many ponies do you see in the image?"
+        )
+        assert result[0]["content"][0]["type"] == "text"
+        assert result[0]["content"][1]["type"] == "image_url"
+        assert result[0]["content"][1]["image_url"] == BLOB_DATA_SUBSTITUTE
+
     def test_does_not_redact_image_url_content_with_non_blobs(self):
         messages = [
             {

_{This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 51ff087. Configure here.}