🤖 AgentStudio - Automate your browser tasks today

AgentStudio transforms a standard computer into an autonomous coding and browsing system. It uses artificial intelligence to plan tasks, write code, run programs, and control your web browser. This tool handles complex web interactions just like a human user would.
🛠 Features
- Autonomous Planning: The agent breaks down large tasks into small, manageable steps.
- Code Generation: It writes and executes code to solve problems.
- Web Automation: The system controls a real Chrome browser instance to interact with websites.
- Local Processing: Your data stays on your machine.
- Simple Setup: Follow one command to install the application.
- Wide Compatibility: It works on both Windows and macOS systems.
💻 Requirements
- Operating System: Windows 10 or Windows 11.
- Memory: At least 8GB of RAM.
- Storage: 2GB of free disk space.
- Internet: A stable connection to retrieve web data and connect to the AI model.
- API Key: You need a free NVIDIA NIM API key to power the agent.
💾 Installation
Follow these steps to set up AgentStudio on your computer.
- Visit the official download page to access the installer file.
- Look for the latest version under the “Assets” section.
- Click the file ending in
.exe to download it to your computer.
- Open the folder where you saved the file.
- Double-click the installer file.
- Follow the prompts on your screen to complete the installation process.
🚀 Running the Agent
Once you install the software, you must link it to your AI provider.
- Open the AgentStudio application from your Start menu or desktop shortcut.
- The program will open a settings window.
- Paste your NVIDIA NIM API key into the designated field.
- Save your settings.
- Type your first instruction into the main interface. For example, you can tell the agent to “Find the price of a laptop on a electronics website.”
- Press the Enter key or click the Run button to start the process.
- Watch the screen as the agent opens the browser, types inputs, and navigates through pages to fulfill your request.
📂 Understanding the Agent Process
The agent functions through a loop. First, it analyzes your request. It decides which tools it needs to complete your goal. It creates a plan and then translates that plan into actions. If you ask the agent to search for information, it uses the Chrome browser to perform the search. It reads the results and decides if it needs to visit a specific link. It continues these actions until it reaches the final result.
The agent makes decisions for each step. You can watch this in the output window. The window logs each move the agent takes. If the agent makes a mistake, it uses its own logic to correct the path and proceed.
⚙️ Advanced Settings
You can modify how the agent behaves in the options menu.
- Browser Visibility: You can choose to hide the browser window to save system resources or keep it visible to watch the automation in real-time.
- Model Selection: Change the internal AI settings if you have access to different model endpoints.
- Logging: You can enable detailed logs if you want to track every function call the agent makes. This helps if you ever need to troubleshoot why an action failed.
💡 Usage Tips
Keep your tasks specific for the best results. Instead of a general request like “Find me a computer,” try “Search for the latest model of Laptop X on Amazon and tell me the current price.” The more context you provide, the better the agent performs.
Ensure your internet connection remains stable during long tasks. If the agent loses connection to the AI provider, the process stops until the connection resumes. You can safely close the agent at any time, but be aware that any currently running task will stop immediately.
🛠 Troubleshooting
- Installation Error: If the installation stops, ensure you have administrative rights on your user account. Try restarting your computer before you try again.
- API Key Failure: Verify that your NVIDIA NIM API key is valid. Check the spelling of the key and ensure there are no extra spaces before or after the code string.
- Browser Does Not Open: Check that you have the latest version of Chrome installed on your machine. The agent relies on the standard installation of Chrome to execute web tasks.
- Slow Response: AI responses rely on your network speed. If the agent appears to hang, please wait a minute as it may be waiting for a response from the model server.