Skip to content

Meta Planner Agent Inference

The Meta Planner Agent inference setup is similar to the Meta Agent Inference, with the addition of Planner Agent capabilities for advanced planning and execution.

It provides a chat interface where you can interact with the onboarded agent and observe the steps it takes to answer your queries.

Meta Planner Agent Interface

The Meta Planner Agent interface allows you to select the Meta Planner Agent template, model, and agent name to begin interacting with the agent. This provides a comprehensive chat environment for advanced planning and execution tasks.

Plan Verifier

The Plan Verifier option allows users to review and validate the plan generated by the Meta Planner Agent before execution. When enabled, the agent displays a detailed plan outlining the steps it will take to answer your query.

After reviewing the plan, you have two options:

  • Approve the Plan: Use the approval option to confirm the plan. The agent will then execute the plan and generate a response to your query.
  • Request a Revised Plan: If you are not satisfied with the plan, you can request a revision. The agent will generate a revised plan based on your feedback, allowing you to review and approve the new plan before execution.

Agent Verifier

The Agent Verifier introduces the Agent Interrupt feature, enabling users to monitor and control agent executions during meta planner agent interactions. This ensures that each step taken by the agent aligns with expected behavior and requirements, providing greater transparency and oversight.

Inference Results

The inference results showcase the chat interaction process. The Meta Planner Agent combines the orchestration capabilities of the Meta Agent with planning features from the Planner Agent, providing comprehensive responses to user queries.

Steps Taken by the Agent

  • You can view the steps taken by the agent to answer your query through the Execution Steps feature.
  • These steps reveal the agents the meta planner agent calls based on the user query, as well as the planning steps, providing transparency into the decision-making process.