Tool Integrations
⚙Tool
Stagehand Agent
Autonomous browser agent for complex multi-step web tasks
Stagehand Agent
The Stagehand Agent is an autonomous browser automation agent. Unlike Stagehand (which executes one action at a time), the Agent plans and executes multi-step browser tasks independently — navigating, interacting, and extracting data with minimal guidance.
Overview
| Property | Value |
|---|---|
| Type | stagehand_agent |
| Category | Tool — Web Scraping |
| Auth | API Key |
Configuration
| Setting | Type | Description |
|---|---|---|
| API Key | Password | Stagehand API key |
| Task | Long text | High-level task description |
| Start URL | Short input | Starting URL |
| Max Steps | Slider | Maximum actions to take |
Outputs
| Field | Type | Description |
|---|---|---|
content | string | Task results |
steps | json | Actions taken |
success | boolean | Task completion status |
Example: Price Comparison
Workflow:
[Starter: Product name] → [Stagehand Agent] → [Agent: Compare] → [Response]Task: Search for "{{starter.input}}" on Amazon and Best Buy. Extract the product name, price, rating, and availability from the first result on each site.
The Agent autonomously navigates both sites, searches, and extracts pricing data.
Tips
- High-level instructions work best — describe the goal, not individual steps
- Set max steps to prevent infinite loops on complex sites
- More expensive than Stagehand — use for genuinely complex multi-step tasks