Deployments

Deployments in Voker represent saved configurations that you can execute via the API.

Run by Deployment Id

a11y.sectionLink Run by Deployment Id

The simplest way to run a deployment is by using its unique deployment ID, which is generated when you create a deployment. When you create a deployment through the Voker UI, you will receive a cURL command along with equivalent commands in Python and JavaScript. These commands will include the Deployment ID and placeholders for any required inputs.

example.bash

API_TOKEN="..."
DEPLOYMENT_ID="..."

curl https://api.voker.ai/api/v1/deployments/$DEPLOYMENT_ID/run \
    --request POST \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $API_TOKEN" \
    --data '{
        "input": {
            "message": "Hello, this is a test message."
        }
    }'

Run Live Graph Version

a11y.sectionLink Run Live Graph Version

You can designate a deployment as Live in the Voker UI, making it the default version. This means that any request to run a deployment without specifying an ID or name will execute the live version. Use the following cURL command to run the currently live deployment:

example.bash

API_TOKEN="..."
VOKER_ID="..."

curl https://api.voker.ai/api/v1/graphs/$VOKER_ID/run \
    --request POST \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $API_TOKEN" \
    --data '{
        "input": {
            "message": "Hello, this is a test message"
        }
    }'

Under your created deployment you will see a “live toggle”

When clicked you will see a label next to the deployment indicating that it is live.

Run by Deployment Name

a11y.sectionLink Run by Deployment Name

Alternatively, you can run a deployment using its name. The deployment name is set when you create it and can be changed at any time.

example.bash

API_TOKEN="..."
VOKER_ID="..."
DEPLOYMENT_NAME="..."

curl https://api.voker.ai/api/v1/graphs/$VOKER_ID/run?v=$DEPLOYMENT_NAME \
    --request POST \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $API_TOKEN" \
    --data '{
        "input": {
            "message": "Hello, this is a test message."
        }
    }'

Caution

If you change the deployment name through the UI, you must update it in your API requests; otherwise, they will fail.

Changing the live deployment will automatically update which version is executed when making this request.

Maintaining Context

a11y.sectionLink Maintaining Context

To maintain context across multiple Voker calls—such as in a chat—you can use the conversation_id parameter.

A successful Voker response includes a conversation_id for follow-up requests:

{
    "result": ... // the result of the Voker call in the format you requested
    "conversation_id": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx" // the conversation ID to use in follow up calls
}

Include this conversation_id in subsequent requests to maintain conversation context.

Example Request With Conversation ID

a11y.sectionLink Example Request With Conversation ID

example.bash

API_TOKEN="..."
DEPLOYMENT_ID="..."
CONV_ID="xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"

curl https://api.voker.ai/api/v1/deployments/$DEPLOYMENT_ID/run \
    --request POST \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $API_TOKEN" \
    --data '{
        "input": {
            "message": "Hello, this is a test message."
        },
        "conversation_id": "'"$CONV_ID"'"
    }'

Full Example in Python

a11y.sectionLink Full Example in Python

example.py

import requests


API_TOKEN = "..."
DEPLOYMENT_ID = "..."

# Initial request (no conversation ID)
response1 = requests.post(
    f"https://api.voker.ai/api/v1/deployments/{DEPLOYMENT_ID}/run",
    headers={"Authorization": f"Bearer {API_TOKEN}"},
    json={
        "input": {
            "message":  "What are some good restaurants in the area?",
        },
    },
)

response1_data = response1.json()
conversation_id = response1_data["conversation_id"]
print(response1_data["result"])

# Follow-up request with conversation ID
response2 = requests.post(
    f"https://api.voker.ai/api/v1/deployments/{DEPLOYMENT_ID}/run",
    headers={"Authorization": f"Bearer {API_TOKEN}"},
    json={
        "input": {
            "message":  "Can you suggest some budget-friendly options?",
        },
        "conversation_id": conversation_id,
    },
)

response2_data = response2.json()
print(response2_data["result"])

Image inputs

a11y.sectionLink Image inputs

The following examples show how to call a Voker with an image input (called img) in Javascript, Python and using cURL.

PNG

a11y.sectionLink PNG

• bash
• python
• javaScript

example.bash

API_TOKEN="..."
FILE_PATH="..."
DEPLOYMENT_ID="..."

PNG_ENCODED=$(base64 $FILE_PATH | tr -d '\n')

curl https://api.voker.ai/api/v1/deployments/$DEPLOYMENT_ID/run \
    --request POST \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $API_TOKEN" \
    --data '{
        "input": {
            "img": "data:image/png;base64,'"$PNG_ENCODED"'"
        }
    }'

example.py

import base64

import requests


API_TOKEN = "..."
FILE_PATH = "..."
DEPLOYMENT_ID = "..."

with open(FILE_PATH, "rb") as image_file:
    png_encoded = base64.b64encode(image_file.read()).decode("utf-8")

response = requests.post(
    f"https://api.voker.ai/api/v1/deployments/{DEPLOYMENT_ID}/run",
    headers={
        "Authorization": f"Bearer {API_TOKEN}",
    },
    json={
        "input": {
            "img": f"data:image/png;base64,{png_encoded}",
        },
    },
)

print(response.json()["result"])

example.js

import * as fs from 'fs';


const API_TOKEN = '...';
const FILE_PATH = '...';
const DEPLOYMENT_ID = '...';

const pngEncoded = fs.readFileSync(FILE_PATH, { encoding: 'base64' });

fetch(
    `https://api.voker.ai/api/v1/deployments/${DEPLOYMENT_ID}/run`,
    {
        method: 'POST',
        headers: {
            'Content-Type': 'application/json',
            Authorization: `Bearer ${API_TOKEN}`,
        },
        body: JSON.stringify({
            input: {
                img: `data:image/png;base64,${pngEncoded}`,
            },
        }),
    },
)
    .then((response) => response.json())
    .then((data) => console.log(data.result));

JPG

a11y.sectionLink JPG

• bash
• python
• javaScript

example.bash

API_TOKEN="..."
FILE_PATH="..."
DEPLOYMENT_ID="..."

JPG_ENCODED=$(base64 $FILE_PATH | tr -d '\n')

curl https://api.voker.ai/api/v1/deployments/$DEPLOYMENT_ID/run \
    --request POST \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $API_TOKEN" \
    --data '{
        "input": {
            "img": "data:image/jpg;base64,'"$JPG_ENCODED"'"
        }
    }'

example.py

import base64

import requests


API_TOKEN = "..."
FILE_PATH = "..."
DEPLOYMENT_ID = "..."

with open(FILE_PATH, "rb") as image_file:
    jpg_encoded = base64.b64encode(image_file.read()).decode("utf-8")

response = requests.post(
    f"https://api.voker.ai/api/v1/deployments/{DEPLOYMENT_ID}/run",
    headers={
        "Authorization": f"Bearer {API_TOKEN}",
    },
    json={
        "input": {
            "img": f"data:image/jpg;base64,{jpg_encoded}",
        },
    },
)

print(response.json()["result"])

example.js

import * as fs from 'fs';


const API_TOKEN = '...';
const FILE_PATH = '...';
const DEPLOYMENT_ID = '...';

const jpgEncoded = fs.readFileSync(FILE_PATH, { encoding: 'base64' });

fetch(
    `https://api.voker.ai/api/v1/deployments/${DEPLOYMENT_ID}/run`,
    {
        method: 'POST',
        headers: {
            'Content-Type': 'application/json',
            Authorization: `Bearer ${API_TOKEN}`,
        },
        body: JSON.stringify({
            input: {
                img: `data:image/jpg;base64,${jpgEncoded}`,
            },
        }),
    },
)
    .then((response) => response.json())
    .then((data) => console.log(data.result));