Deploy Your Agents

This guide shows you how to deploy AI Agents with the videosdk-agents python package. Once your AI Agent is ready to use, you need to create an AI Worker. The AI Worker is responsible for running your AI Agent. Before proceeding, ensure you have completed the steps under Prerequisites.

Prerequisites

To deploy your AI Worker, make sure you have:

• Created an AI Worker using the Create AI Worker API.
• A VideoSDK authentication token (generate from VideoSDK Dashboard)

YAML Configuration

Create a videosdk.yaml file with the following structure:

version: "1.0"

worker:
  id: your_worker_id
  entry:
    path: entry_point_for_worker

env: # Optional to run your agent locally
  path: "./.env"

secrets:
  VIDEOSDK_AUTH_TOKEN: your_auth_token

deploy:
  cloud: true

Field Descriptions

Field	Description
worker.id	The workerId obtained from the Create AI Worker API.
worker.entry.path	Path to the entry point script for your AI Worker.
env.path	Path to your .env file, used only when running the agent locally.
secrets.VIDEOSDK_AUTH_TOKEN	Your VideoSDK auth token (required for deployment).
deploy.cloud	Set to true to allow deploying the worker to VideoSDK Cloud, when using the deploy command. Use false to avoid accidental deploys.

CLI Commands

• Run the AI Worker locally for Testing.

videosdk.agents run 

• Deploy the AI Worker.

videosdk.agents deploy

Next Steps

After deploying your AI Worker, you can start using it by:

1. Creating a new session using the Start Session API
2. Ending the session using the End Session API