Skip to content

Meta Agent Inference

The Meta Agent inference setup provides a comprehensive chat interface where you can interact with onboarded agents and observe their decision-making processes in real-time. This system allows you to test and evaluate agent performance while gaining insights into their reasoning steps.

Getting Started

To begin using the Meta Agent inference system:

  1. Select Configuration: Choose your Meta Agent template from the available options
  2. Choose Model: Select the appropriate AI model for your use case
  3. Pick Agent: Select the specific agent you want to interact with from your onboarded agents
  4. Start Chatting: Begin your conversation with the selected agent

The interface provides a clean, intuitive chat experience similar to popular messaging applications, making it easy to communicate with your AI agents.

Understanding Agent Responses

When you submit a query to the Meta Agent, it processes your request through multiple layers:

  • Query Analysis: The agent first analyzes your question to understand the context and requirements
  • Agent Selection: Based on the query type, the Meta Agent determines which specialized sub-agents to invoke
  • Collaborative Processing: Multiple agents may work together to provide comprehensive answers
  • Response Synthesis: The Meta Agent combines insights from various sources to deliver a coherent response

Viewing Agent Steps

The system provides complete transparency into the agent's decision-making process:

  • Step-by-Step Breakdown: Access detailed information about each step the agent takes
  • Agent Routing: See which specific sub-agents were called and why
  • Decision Logic: Understand the reasoning behind agent selections
  • Processing Timeline: View the sequence of operations performed

This transparency feature is crucial for debugging, optimization, and building trust in the agent's capabilities.

Chat History Management

The platform includes robust chat history functionality:

  • Session Persistence: All conversations are automatically saved
  • Easy Retrieval: Access previous conversations through the "Old Chats" dropdown
  • Search Capability: Quickly find specific conversations or topics
  • Reference Material: Use past chats for analysis, training, or documentation purposes

This feature enables continuous learning and improvement of your interaction patterns with the agents.