orgo 0.0.14__tar.gz → 0.0.16__tar.gz

{orgo-0.0.14 → orgo-0.0.16}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: orgo
-Version: 0.0.14
+Version: 0.0.16
 Summary: Computers for AI agents
 Author: Orgo Team
 License: MIT

{orgo-0.0.14 → orgo-0.0.16}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "orgo"
-version = "0.0.14"
+version = "0.0.16"
 description = "Computers for AI agents"
 authors = [{name = "Orgo Team"}]
 license = {text = "MIT"}

{orgo-0.0.14 → orgo-0.0.16}/src/orgo/computer.py

@@ -4,8 +4,9 @@ import os
 import io
 import base64
 import logging
-from typing import Dict, List, Any, Optional, Callable, Union
+from typing import Dict, List, Any, Optional, Callable
 from PIL import Image
+import requests
 from requests.exceptions import RequestException
 
 from .api.client import ApiClient
@@ -97,13 +98,33 @@ class Computer:
     def screenshot(self) -> Image.Image:
         """Capture screenshot and return as PIL Image"""
         response = self.api.get_screenshot(self.project_id)
-        img_data = base64.b64decode(response.get("image", ""))
-        return Image.open(io.BytesIO(img_data))
+        image_data = response.get("image", "")
+        
+        # Check if it's a URL (new format) or base64 (legacy format)
+        if image_data.startswith(('http://', 'https://')):
+            # Download image from URL
+            img_response = requests.get(image_data)
+            img_response.raise_for_status()
+            return Image.open(io.BytesIO(img_response.content))
+        else:
+            # Legacy base64 format
+            img_data = base64.b64decode(image_data)
+            return Image.open(io.BytesIO(img_data))
     
     def screenshot_base64(self) -> str:
         """Capture screenshot and return as base64 string"""
         response = self.api.get_screenshot(self.project_id)
-        return response.get("image", "")
+        image_data = response.get("image", "")
+        
+        # Check if it's a URL (new format) or base64 (legacy format)
+        if image_data.startswith(('http://', 'https://')):
+            # Download image from URL and convert to base64
+            img_response = requests.get(image_data)
+            img_response.raise_for_status()
+            return base64.b64encode(img_response.content).decode('utf-8')
+        else:
+            # Already base64
+            return image_data
     
     # Execution methods
     def bash(self, command: str) -> str:

{orgo-0.0.14 → orgo-0.0.16}/src/orgo/prompt.py

@@ -323,12 +323,28 @@ class AnthropicProvider:
                 response = api_client.get_screenshot(computer_id)
                 if callback:
                     callback("tool_result", {"type": "image", "action": "screenshot"})
+                
+                # The API now returns a URL instead of base64 data
+                # We need to fetch the image from the URL and convert it to base64
+                image_url = response.get("image", "")
+                
+                if not image_url:
+                    raise ValueError("No image URL received from API")
+                
+                # Fetch the image from the URL
+                import requests
+                img_response = requests.get(image_url)
+                img_response.raise_for_status()
+                
+                # Convert to base64
+                image_base64 = base64.b64encode(img_response.content).decode('utf-8')
+                
                 return {
                     "type": "image",
                     "source": {
                         "type": "base64",
                         "media_type": "image/jpeg",
-                        "data": response.get("image", "")
+                        "data": image_base64
                     }
                 }
                 

{orgo-0.0.14 → orgo-0.0.16}/src/orgo.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: orgo
-Version: 0.0.14
+Version: 0.0.16
 Summary: Computers for AI agents
 Author: Orgo Team
 License: MIT