​
Using Vertex AI with Flowstack

Vertex AI can be integrated with Flowstack using various methods including REST API, Python, and Node.js. Below are the implementations for each.

​
1. Using REST API

To make a request using the REST API, use the following Curl command. Replace YOUR_GCP_PROJECT_ID, MODEL_ID, and YOUR_GCLOUD_ACCESS_TOKEN with your Google Cloud project ID, Vertex AI model ID, and GCloud access token, respectively.

curl -X POST "https://vertex-ai.flowstack.ai/v1/projects/YOUR_GCP_PROJECT_ID/locations/us-central1/publishers/google/models/MODEL_ID:predict" \
-H "Authorization: Bearer YOUR_GCLOUD_ACCESS_TOKEN" \
-H "Content-Type: application/json" \
-H "Flowstack-Auth: Bearer YOUR_FLOWSTACK_KEY" \
-d '{
    "instances": [
        {
            "prompt": "Hello"
        }
    ],
    "parameters": {
        "temperature": 0,
        "maxOutputTokens": 5,
        "topP": 0,
        "topK": 1
    }
}'

​
2. Using Python

For Python integration, use the following script. Ensure to replace YOUR_GCP_PROJECT_ID, MODEL_ID, YOUR_GCLOUD_ACCESS_TOKEN, and YOUR_FLOWSTACK_KEY with your actual credentials.

import requests
import json

url = f"https://vertex-ai.flowstack.ai/v1/projects/YOUR_GCP_PROJECT_ID/locations/us-central1/publishers/google/models/MODEL_ID:predict"

headers = {
    "Authorization": f"Bearer YOUR_GCLOUD_ACCESS_TOKEN",
    "Content-Type": "application/json",
    "Flowstack-Auth": "Bearer YOUR_FLOWSTACK_KEY"
}

data = {
    "instances": [
        {
            "prompt": "Hello"
        }
    ],
    "parameters": {
        "temperature": 0,
        "maxOutputTokens": 5,
        "topP": 0,
        "topK": 1
    }
}

response = requests.post(url, headers=headers, data=json.dumps(data))
print(response.text)

​
3. Using Node.js

For Node.js, the following implementation can be used. Replace YOUR_GCP_PROJECT_ID, MODEL_ID, YOUR_GCLOUD_ACCESS_TOKEN, and YOUR_FLOWSTACK_KEY with the respective credentials.

const axios = require('axios');

const url = `https://vertex-ai.flowstack.ai/v1/projects/YOUR_GCP_PROJECT_ID/locations/us-central1/publishers/google/models/MODEL_ID:predict`;

const headers = {
    "Authorization": `Bearer YOUR_GCLOUD_ACCESS_TOKEN`,
    "Content-Type": "application/json",
    "Flowstack-Auth": `Bearer YOUR_FLOWSTACK_KEY`
};

const data = {
    instances: [
        {
            prompt: "Hello"
        }
    ],
    parameters: {
        temperature: 0,
        maxOutputTokens: 5,
        topP: 0,
        topK: 1
    }
};

axios.post(url, data, { headers: headers })
    .then(response => {
        console.log(response.data);
    })
    .catch(error => {
        console.error('Error:', error);
    });
AWS BedrockMistral AI