Browser AI Agent
AI agent that remotely controls browsers using vision-based automationProperties
| Property | Value |
|---|---|
| Type | browser-ai-agent |
| Family | AI |
| Group | Browser |
Parameters
| Name | Label | Type | Description | Required |
|---|---|---|---|---|
| model | Model | Select | AI model to use for vision-based automation | Yes |
| task | Browser Task | String | The task to perform in the browser (e.g., ‘Click the Sign In button’) | Yes |
| systemPrompt | System Prompt | String | System prompt for the AI agent | No |
| maxSteps | Max Steps | Number | Maximum number of automation steps to take | No |
| temperature | Temperature | Number | AI creativity level (lower = more deterministic) | No |
Outputs
| Name | Label | Type | Description |
|---|---|---|---|
| success | success | boolean | Whether the task was completed successfully |
| stepsExecuted | stepsExecuted | number | Number of steps executed |
| actionLog | actionLog | array | Log of all actions taken with screenshots |
| finalScreenshot | finalScreenshot | string | URL of the final screenshot |
| htmlReport | htmlReport | string | HTML report with step-by-step screenshots |
| animationScripts | animationScripts | array | Animation data for video playback |