OnDemand + FULLY FREE 3.7 Sonnet & O4-mini: This is the EASIEST Way to MAKE AI AGENTS!

Key Concepts

AI Agents, AI Automation, On-Demand AI Operating System, Agent Marketplace, REST API Agent, Knowledge-Based Agent, IoT Agent, Agent Flow Builder, LLM Block, Chatbot Creator, Model Selection (GPT-4.1, Claude 3.7 Sonnet, Gemini 2.5 Pro), API Integration, Cron Jobs, Webhooks, Bring Your Own Model.

On-Demand AI Platform Overview

The video introduces On-Demand, an AI operating system designed for building and deploying custom AI agents and automation workflows. The platform offers $50 in free credits to new users who sign up through the link in the description, allowing them to experiment with agent creation and testing without upfront costs. Users can also monetize their agents by sharing them on the platform's agent marketplace.

Agent Marketplace and Pre-Made Agents

The platform features an agent marketplace with a variety of pre-made agents, including those for perplexity, coding, and finance. Users can select and use these agents directly or create their own and share them on the marketplace.

Creating Custom AI Agents

The video details the process of creating different types of AI agents:

1. Knowledge-Based Agent

• Process:
  1. Navigate to the agent marketplace and select the option to create a new agent.
  2. Choose the "Knowledge-Based Agent" type.
  3. Enter agent details such as name, icon, and category (e.g., general, finance).
  4. Provide an agent description and optional conversation starters (starter prompts).
  5. Upload a knowledge base in formats like PDF or CSV.
  6. Configure privacy policy settings.
  7. Save the agent.
• Usage: The agent can be published to the marketplace or used directly in the playground. In the playground, the model treats the knowledge base as a tool it can call upon when responding to prompts.

2. REST API Agent

• Process:
  1. Select the "REST API Agent" type when creating a new agent.
  2. Enter agent details and configuration fields for request headers (e.g., authorization, API key).
  3. Input the API schema, using provided examples as a guide (e.g., air quality API).
  4. Deploy the agent.
• Usage: The agent can be used in the playground after deployment.

Agent Flow Builder

The agent flow builder allows users to create custom automation workflows using AI models.

• Process:
  1. Create a new workflow in the agent flow builder.
  2. Configure a trigger, such as an API request or a cron job for automated execution at specified intervals.
  3. Set the model and prompt, along with the agent to be used.
  4. Connect the output to various destinations, including email, phone number, Slack, or custom webhooks.
• Features:
  • LLM Block: Allows integrating custom agents into the workflow.
  • Voice Mode Agents: Enables the creation of advanced voice-activated agents.

Model Selection and Settings

The platform offers a range of models, including GPT-4.1, Claude 3.7 Sonnet, and Gemini 2.5 Pro, which are available under the free credits. Users can select the desired model in the playground and adjust settings such as reasoning effort. Selecting the "state-of-the-art" model automatically switches to Gemini 2.5 Pro.

API Integration

The platform allows users to integrate AI agents into their own applications via an API.

• Process:
  1. Select the code icon in the playground.
  2. Choose the desired programming language.
  3. The platform generates the code required for integration.

Bring Your Own Model

Users can integrate custom models from platforms like Hugging Face for greater customization of workflows.

Chatbot Creator

The platform includes a chatbot creator that allows users to build chatbots with selected models and integrate their created agents for enhanced functionality.

Conclusion

On-Demand provides a comprehensive platform for building, deploying, and monetizing AI agents and automation workflows. With its user-friendly interface, diverse agent types, and integration capabilities, it empowers users to create custom AI solutions tailored to their specific needs. The availability of free credits and a wide range of models makes it an accessible option for both beginners and experienced AI developers.