Vertex AI Integration Skill
Embed GenAI into Google Cloud Application Integrations.
Advanced References
- •Predict API Patterns: See vertex-ai-predict.md
Overview
Use Vertex AI tasks to add AI capabilities to your integration workflows.
Supported Models
- •
text-bison- Text generation - •
chat-bison- Conversational AI - •
gemini-pro- Advanced reasoning
Integration Flow
text
API Trigger → Set Parameters → Vertex AI Predict → Map Response
Sample Configuration
Prompt Request Template
json
{
"instances": [{
"prompt": "$TextPrompt$"
}],
"parameters": {
"temperature": 0.2,
"maxOutputTokens": 768,
"topP": 0.8,
"topK": 40
}
}
Required Variables
| Variable | Type | Description |
|---|---|---|
TextPrompt | String | Input prompt |
ModelId | String | Model name (e.g., text-bison@001) |
ProjectId | String | GCP Project ID |
Region | String | GCP Region (e.g., us-central1) |
Setup Steps
- •Enable APIs: Vertex AI API, Application Integration API
- •Create Service Account: Grant
Vertex AI Userrole - •Create Integration: Add Vertex AI Predict task
- •Configure Authentication: Attach service account
- •Test: Run integration with sample prompts
Endpoint Format
bash
publishers/google/models/{ModelId}
Response Handling
javascript
// Extract content from predictions const content = response.predictions[0].content;