Agent5ive LogoAgent5ive Docs
Visit Platform

Getting Started with Agent5ive

Learn the basics of creating and using AI agents with Agent5ive.

Prerequisites

Before you begin using Agent5ive, make sure you have:

  • A modern web browser (Chrome, Firefox, Safari, or Edge)
  • An account on the Agent5ive platform
  • Basic understanding of what you want your agent to accomplish

Creating Your First Agent

1

Sign in to Agent5ive

Navigate to and sign https://agent5ive.com in with your account credentials.

2

Start a new agent

Click on the 'Start Building Your Agent' button on the homepage or the 'New Agent' button in your dashboard.

3

Describe your agent

Use natural language to describe what you want your agent to do. Be specific about its purpose, tasks, and any constraints.

4

Enhance with tools

Select from available tools to enhance your agent's capabilities. These tools include Web Search, News Search, Google Maps, and other search-related tools.

5

Generate your agent

Click the 'Generate' button to create your agent. You'll see an overview of your agent including System Prompt, Instructions, Capabilities, and main characteristics.

6

Deploy your agent

Click the 'Deploy' button to deploy your agent privately. This step is required before you can test your agent.

7

Test your agent

Interact with your deployed agent in the testing environment to ensure it behaves as expected. Make adjustments as needed.

8

Publish your agent

When you're satisfied with your agent's performance, publish it publicly to make it available for integration via SDK, iframe, WhatsApp, or Telegram.

Using Sample Agent Templates

To speed up the agent creation process, Agent5ive offers pre-made templates for common agent types. When you start creating a new agent, you'll see a "Sample Agent Prompts" option that provides templates for:

Sales AgentSales Agent

A persuasive agent that helps promote products, handle objections, and close sales with a professional approach.

Customer SupportCustomer Support

A helpful agent that answers product questions, resolves issues, and ensures customer satisfaction.

Content WriterContent Writer

A creative agent that generates blog posts, articles, and marketing copy tailored to your brand voice.

Interview CoachInterview Coach

A supportive agent that helps prepare for job interviews with practice questions and feedback.

Travel PlannerTravel Planner

An informative agent that creates personalized travel itineraries, recommends attractions, and offers travel tips.

Dating CoachAli G Dating Coach

A humorous agent with Ali G's personality that offers dating advice and conversation starters with comedic flair.

Confidence BuilderArnold Confidence Builder

A motivational agent with Arnold's persona that provides encouragement, workout tips, and confidence-boosting advice.

These templates provide a great starting point that you can then customize to your specific needs.

Example: Creating a Research Assistant Agent

Let's walk through an example of creating a research assistant agent that can help gather information on specific topics.

Step 1: Describe your agent

"I want to create a research assistant agent that can help me gather information on various topics. The agent should be able to search the web, summarize information, and provide sources for its findings. It should be helpful, accurate, and able to understand complex research queries."

Step 2: Enhance with tools

  • Web Search Tool - Allows the agent to search the internet for information
  • News Search Tool - Helps the agent find recent news articles on topics
  • Google Maps Search - Enables the agent to find locations and geographic information

Step 3: Generate your agent

After clicking the Generate button, you'll see an overview of your agent including its System Prompt, Instructions, Capabilities, and main characteristics.

Step 4: Deploy your agent

Click the Deploy button to deploy your agent privately. This step is required before you can test your agent.

Step 5: Test

You might test your deployed agent with queries like:

"Find me recent research on climate change mitigation strategies"

Step 6: Publish your agent

Based on the test results, you might refine your agent before publishing it publicly. Once published, you can integrate it via SDK, iframe, WhatsApp, or other available methods.

Iterating on Your Agents

Agent5ive makes it easy to iterate and improve your agents over time. There are two main ways to refine your agents:

Direct Modifications

You can directly modify certain aspects of your agent after creation:

  • System prompt
  • Agent capabilities and limitations
  • Agent name and description
  • Advanced settings (temperature, etc.)

These changes can be made in the Configuration & Preview tab after your agent is generated but before it's deployed.

"Start New Chat Based on Vision"

For changes that require a new agent (like changing tools or agent-as-tools):

  1. In any chat with your agent, click "Start new chat based on vision"
  2. This preserves the agent's core purpose and personality
  3. You can now select different tools, voice settings, etc.
  4. Generate & Deploy the new agent with your updated configuration

This feature is perfect for evolving your agents without starting from scratch.

When to Use "Start New Chat Based on Vision"

  • Adding or removing tools - When you want to change which tools your agent can use
  • Changing agent-as-tools - When you want to modify which other agents your agent can call
  • Enabling/disabling voice - When you want to add or remove voice capabilities
  • Enabling/disabling image vision - When you want to add or remove image processing capabilities
  • Creating variations - When you want to create similar agents with different specializations

⚠️ New Experimental Feature: Advanced Tools (MCP)

Agent5ive now supports connecting to external tool servers using the Model Context Protocol (MCP). This experimental feature allows your agents to access tools from services like Google Calendar, Gmail, GitHub, and more through Composio's MCP servers.

To learn more about this feature, visit our Tool Integration documentation.