Skip to main content

My Agents

The My Agents page is your central hub for viewing, testing, and managing all your AI agents. Access it from the sidebar to see your complete agent inventory.

Agent List​

All your agents are displayed in a table format showing key configuration details at a glance.

Columns​

NAME & PROMPT
Agent name with a preview of the system prompt.

PROVIDER
The AI provider (OpenAI, Anthropic, Google, etc.).

MODEL
Specific model powering the agent (GPT-4, Claude 3.5 Sonnet, etc.).

TEMPERATURE
Creativity setting (0.0 = consistent, 2.0 = creative).

CHAT MODE
Shows "Enabled" if Interactive Mode is active, allowing conversational memory.

ACTIONS
Quick action buttons for each agent.

Page Actions​

Refresh
Update the agent list to see latest changes.

Create with AI
Launch the AI-assisted agent builder for quick setup.

+ New Agent
Open the Creation Wizard to build a new agent manually.

Agent Actions​

Click the three-dot menu (⋮) next to any agent to access these options:

Chat​

Opens the Interactive Chat interface to test your agent in real-time.

Chat Interface:

  • Message Input - Type queries at the bottom
  • Clear Chat - Reset conversation history
  • Interactive/Auto Run Toggle - Switch between manual and automated modes
  • Conversation Display - View full chat history with your agent
Testing Agents

The chat interface is perfect for validating system prompts, testing tool integrations, and refining agent behavior before deployment.

Edit​

Opens the Creation Wizard pre-filled with your agent's current configuration. Make changes to any step and save to update.

You can modify:

  • System prompt and context
  • Model selection and advanced settings
  • Attached tools
  • RAG knowledge files
  • Scheduling settings

View Logs​

Opens the Agent Runs page showing execution history and performance metrics.

Displayed Metrics:

  • Total Tokens - Cumulative token usage across all runs
  • Total Cost - Combined cost of all executions
  • Successful Runs - Number of completed executions
  • Failed Runs - Number of errors or failures

Search Runs
Use the search bar to filter executions by date, status, or keywords.

info

Logs help you debug issues, track performance, and optimize agent configurations.

Delete​

Permanently removes the agent from your workspace.

Irreversible Action

Deleting an agent cannot be undone. All associated data, logs, and configurations will be lost.

Agent Modes​

Interactive Mode​

Manual conversation mode where you control each interaction.

When to use:

  • Testing agent responses
  • Customer support chatbots
  • Tutoring or coaching agents
  • Q&A assistants

Button: Blue "Interactive" button in chat interface

Auto Run Mode​

Automated execution mode where the agent runs on schedule or triggers.

When to use:

  • Scheduled reports
  • Background monitoring
  • Automated workflows
  • Batch processing

Button: Purple "Auto Run" button in chat interface

Toggle anytime by clicking the mode button in the chat interface.

Status Notifications​

VibeAgent displays real-time notifications for:

  • Agent status changes (Interactive Mode enabled/disabled)
  • Execution status (Run started, completed, failed)
  • Error alerts (Configuration issues, API failures)
  • Save confirmations (Settings updated successfully)

Notifications appear as popups in the top-right corner.

Agent Control Dashboard​

View aggregate analytics across all agents.

Metrics:

  • Total Agents - Active agent count
  • Tools Used - Total tool executions
  • Shortest Run - Fastest agent execution time
  • Avg Runtime - Mean execution time
  • Longest Run - Slowest agent execution time

Charts:

  • Tools per Agent - Distribution of tool usage
  • Tool Usage by Type - Most frequently called tools
  • Runtime vs Tools Used - Performance correlation
Performance Insights

Use dashboard metrics to identify slow agents, optimize tool configurations, and improve overall efficiency.

Quick Tips​

Finding Agents
Look for the preview text under each agent name - it shows the beginning of the system prompt for quick identification.

Quick Testing
Use the chat icon in the Actions column to instantly test an agent without opening menus.

Edit Shortcuts
Click the pencil icon for direct access to the editor.

Monitor Usage
Check "View Logs" regularly to track token consumption and manage costs.

Need to create a new agent? Check out the Creation Wizard guide.