Create Your First Agent

This guide walks you through creating an AI agent, configuring its personality, and starting your first conversation.

Prerequisites

Before creating an agent, make sure you have:

An Aigentic account (Sign Up)
Optionally, an API key for a premium provider (Prerequisites)

Free NVIDIA NIM and Z.AI models are available without any API key.

Step 1: Open the Agent Creator

Open Agents from the sidebar (or click Manage agents on the dashboard) and click the New Agent button.

Step 2: Name Your Agent

Enter a descriptive name for your agent (up to 50 characters). This name appears on the agent card, in chat headers, and in logs. Examples:

“Research Assistant”
“Code Reviewer”
“Daily Digest Bot”

Step 3: Select a Model

Choose the AI model that powers your agent. Models are organized into three tiers:

Tier	Models	Best For
Fast	Claude 3.5 Haiku, GPT-4o Mini, Grok 3 Mini, GLM-4.7 Flash, MiniMax M2.5 Highspeed	Quick responses, simple tasks
Balanced	Claude Sonnet 4, GPT-4o, Grok 4.1 Fast Reasoning, GLM-4.7, Kimi K2 Instruct, DeepSeek V3.1 Terminus	General purpose, coding, analysis
Powerful	Claude Opus 4.6, OpenAI o1, GLM-5, DeepSeek V3.2, Kimi K2.5, MiniMax M2.5	Complex reasoning, creative work

The default model is DeepSeek V3.2 (NVIDIA NIM, free). Free models (NVIDIA NIM, Z.AI GLM-4.7 Flash) are marked as such in the model selector. See Model Selection for detailed comparisons.

Step 4: Pick an Avatar

Choose an avatar for your agent from the avatar picker (blue, gold, green, or red). This is the icon that represents your agent on the Agents page and in the profile dialog. You can change it later.

Step 5: Write a System Prompt (Optional)

The Personality (SOUL.md) textarea defines your agent’s personality, behavior, and instructions. It shapes every response the agent generates. You can leave it blank and add a personality later using the full Personality Editor, which includes starter templates.

Here is a simple system prompt to get started:

# Research Assistant

You are a helpful research assistant. Your job is to help the user find, summarize, and analyze information.

## Personality
- Be concise and clear
- Cite sources when possible
- Ask clarifying questions when the request is ambiguous

## Guidelines
- Focus on accuracy over speed
- If you are unsure about something, say so
- Format responses with headers and bullet points for readability

## Constraints
- Do not make up information
- Stay on topic

See Personality & System Prompts for advanced examples and best practices.

Step 6: Create and Start Chatting

Click Create Agent. Your agent is ready immediately. Open Chat from the sidebar (or click the chat icon on the agent card), type a message, and press Enter to send it.

Example First Conversation

You: What are the main differences between REST and GraphQL APIs?

Agent: Here are the key differences between REST and GraphQL:

**Architecture**
- REST uses multiple endpoints (e.g., /users, /posts)
- GraphQL uses a single endpoint with a query language

**Data Fetching**
- REST returns fixed data structures per endpoint
- GraphQL lets the client specify exactly which fields to return

**Over/Under Fetching**
- REST often returns more or less data than needed
- GraphQL eliminates this by letting clients define their data requirements

...

What’s Next

Now that you have a running agent, explore these topics:

Personality & System Prompts — Write more sophisticated personality definitions
Agent Tools — Understand and configure the available tools
Skills — Install skills from the ClawHub catalog or generate them with SkillForge
Channels — Connect your agent to Discord, Telegram, or Slack
Scheduling — Set up recurring tasks for your agent