Scale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.
Rust agentai-agentapplecomputer-usecomputer-use-agentcontainerizationcuadesktop-automationhacktoberfestlumemacosmanusoperatorswiftvirtualizationvirtualization-frameworkwindowswindows-sandbox
[CI] Add Cua VLM Router Models to Test Harness (#931)
Harsh Vermad8189043b840
diff --git a/.github/workflows/ci-test-models.yml b/.github/workflows/ci-test-models.ymlindex e2151c91f2..36898140a1 100644--- a/.github/workflows/ci-test-models.yml+++ b/.github/workflows/ci-test-models.yml@@ -39,6 +39,14 @@ jobs:# Gemini CU Preview- gemini-2.5-computer-use-preview-10-2025+ # Cua VLM Router Models (via inference.cua.ai)+ - cua/anthropic/claude-haiku-4.5+ - cua/anthropic/claude-sonnet-4.5+ - cua/anthropic/claude-opus-4.5+ - cua/microsoft/fara-7b+ - cua/google/gemini-3-flash-preview+ - cua/google/gemini-3-pro-preview+# InternVL# - huggingface-local/OpenGVLab/InternVL3_5-1B# - huggingface-local/OpenGVLab/InternVL3_5-2B@@ -141,6 +149,7 @@ jobs:GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }}OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }}HF_TOKEN: ${{ secrets.HF_TOKEN }}+ CUA_API_KEY: ${{ secrets.CUA_API_KEY }}- name: Calculate test duration and prepare messageif: always()diff --git a/tests/agent_loop_testing/agent_test.py b/tests/agent_loop_testing/agent_test.pyindex 98693959a2..4642d7acbc 100644--- a/tests/agent_loop_testing/agent_test.py+++ b/tests/agent_loop_testing/agent_test.py@@ -11,6 +11,7 @@ import base64import sysfrom io import BytesIOfrom pathlib import Path+from typing import Any, Dict, Optionalfrom PIL import Image, ImageDraw@@ -107,6 +108,21 @@ class MockComputer:async def type_text(self, text: str) -> None:await asyncio.sleep(0.1)+ async def playwright_exec(self, command: str, params: Optional[Dict] = None) -> Dict[str, Any]:+ """Mock playwright execution for BrowserTool compatibility."""+ await asyncio.sleep(0.1)++ # Handle screenshot command - return mock screenshot data+ if command == "screenshot":+ return {"success": True, "screenshot": self._image} # Return the base64 screenshot++ # Handle get_current_url command+ if command == "get_current_url":+ return {"success": True, "url": "desktop://mock"}++ # All other commands just return success+ return {"success": True}+async def test_cua_agent(model_name: str):"""Test Cua agent with mock computer."""