Skip to main content
Version: 2026-05-26

Agents

Agent Management

Agents are configurable AI personas that customize 4C:me's behavior for specific use cases and workflows.

What Are Agents?​

An agent is a preconfigured AI setup that includes:

  • A custom system prompt that defines the agent's behavior and expertise
  • Associated files and documents the agent can reference
  • Depth controls for how deeply the agent explores topics
  • Access permissions defining who can use the agent

Using an Agent​

Agent Selector highlighted

Agent Selector open

  1. In 4C:me, click Select Agent (BETA) in the top bar
  2. Browse or search for the agent you need
  3. Select the agent — it becomes active for your current conversation
  4. The agent's system prompt and context shape all responses
  5. The platform remembers your last-used agent per conversation

Creating an Agent​

Create New Agent

Click Create New Agent in the Agent Management page to start a guided wizard. The wizard walks you through four steps:

Step 1: Agent Details​

Configure the basic identity of your agent:

  • Icon — Choose an emoji to represent the agent
  • Name — A descriptive name for the agent
  • ID — A unique identifier (auto-generated from the name)
  • Description — A brief summary of what the agent does
  • Category — Organize agents into categories
  • User Groups — Control which groups can access the agent
  • Visibility — Toggle between Public (all users) and Private (restricted)

Step 2: Agent Settings​

Agent Settings

Configure the agent's AI behavior:

  • Default Model — Choose which AI model the agent uses by default
  • Model Configuration — Click the expand arrow next to the model selector to reveal advanced model parameters:
    • Temperature — Controls response randomness (lower = more deterministic, higher = more creative)
    • Max tokens — Maximum length of the generated response
    • Top P — Nucleus sampling threshold
    • Frequency penalty — Reduces repetition of frequently used tokens
    • Presence penalty — Encourages the model to explore new topics
    • Top K — Limits token selection to the top K most probable tokens
    • Search Retries — Number of retry attempts for search operations
  • Agent Instructions (System Prompt) — Define the agent's personality, expertise, and instructions using a rich-text editor. This prompt is sent as the system message in every conversation with the agent.

Step 3: Example Prompts​

Example Prompts

Configure how users interact with the agent:

  • Welcome Message — Set a custom greeting users see when they start a conversation with the agent
  • Example Prompts — Add sample prompts that appear below the welcome message. These help users understand what the agent can do and provide quick-start suggestions. Each prompt consists of:
    • A title displayed as a clickable suggestion
    • A message body that gets inserted into the chat when the user clicks the prompt
    • Optional attached documents that are automatically included in the conversation context whenever the user selects the prompt

Click + Add Prompt to create new example prompts.

Step 4: Agent Files​

Agent Files

Manage the agent's knowledge base:

  • Upload documents the agent can reference during conversations
  • Browse and sort attached files
  • Remove files no longer needed

Uploaded files become part of the agent's context, allowing it to answer questions based on their content.

Editing an Agent​

Agent Detail

Click on any agent card in the grid to open the same four-step wizard with the agent's current configuration. Edit any field and save your changes.

Agent Visibility​

Private by Default​

Agents are set to private by default, meaning:

  • Only the creator and explicitly granted users/groups can see the agent
  • This prevents accidental exposure of specialized agents
  • Administrators can change visibility settings

Public Agents​

Public agents are available to all authenticated users on the platform.

tip

Use agents for specialized tasks like code review, document analysis, or domain-specific Q&A. The system prompt ensures consistent, focused responses.