Switch all to jpeg

browser_use/agent/cloud_events.py

@@ -155,7 +155,7 @@ class CreateAgentStepEvent(BaseEvent):
 		# Capture screenshot as base64 data URL if available
 		screenshot_url = None
 		if browser_state_summary.screenshot:
-			screenshot_url = f'data:image/png;base64,{browser_state_summary.screenshot}'
+			screenshot_url = f'data:image/jpeg;base64,{browser_state_summary.screenshot}'
 			import logging
 
 			logger = logging.getLogger(__name__)

browser_use/agent/prompts.py

@@ -367,8 +367,8 @@ Available tabs:
 				content_parts.append(
 					ContentPartImageParam(
 						image_url=ImageURL(
-							url=f'data:image/png;base64,{screenshot}',
-							media_type='image/png',
+							url=f'data:image/jpeg;base64,{screenshot}',
+							media_type='image/jpeg',
 							detail=self.vision_detail_level,
 						),
 					)

browser_use/llm/google/serializer.py

@@ -101,13 +101,13 @@ class GoogleMessageSerializer:
 							# Handle images
 							url = part.image_url.url
 
-							# Format: data:image/png;base64,<data>
+							# Format: data:image/jpeg;base64,<data>
 							header, data = url.split(',', 1)
 							# Decode base64 to bytes
 							image_bytes = base64.b64decode(data)
 
 							# Add image part
-							image_part = Part.from_bytes(data=image_bytes, mime_type='image/png')
+							image_part = Part.from_bytes(data=image_bytes, mime_type='image/jpeg')
 
 							message_parts.append(image_part)
 

browser_use/llm/ollama/serializer.py

@@ -47,7 +47,7 @@ class OllamaMessageSerializer:
 				url = part.image_url.url
 				if url.startswith('data:'):
 					# Handle base64 encoded images
-					# Format: data:image/png;base64,<data>
+					# Format: data:image/jpeg;base64,<data>
 					_, data = url.split(',', 1)
 					# Decode base64 to bytes
 					image_bytes = base64.b64decode(data)

browser_use/llm/tests/test_gemini_image.py

@@ -42,10 +42,10 @@ def create_random_text_image(text: str = 'hello world', width: int = 4000, heigh
 
 	# Convert to base64
 	buffer = io.BytesIO()
-	image.save(buffer, format='PNG')
+	image.save(buffer, format='JPEG')
 	img_data = base64.b64encode(buffer.getvalue()).decode()
 
-	return f'data:image/png;base64,{img_data}'
+	return f'data:image/jpeg;base64,{img_data}'
 
 
 async def test_gemini_image_vision():

examples/features/add_image_context.py

@@ -77,8 +77,8 @@ def create_sample_images() -> list[ContentPartTextParam | ContentPartImageParam]
 			content_parts.append(
 				ContentPartImageParam(
 					image_url=ImageURL(
-						url=f'data:image/png;base64,{item["value"]}',
-						media_type='image/png',
+						url=f'data:image/jpeg;base64,{item["value"]}',
+						media_type='image/jpeg',
 					),
 				)
 			)