Planning Lab

See how an application creates an explicit multi-step plan, tracks step progress, and revises remaining work when an observation makes the current plan inappropriate.

GuidedNo API keys in LabAI Agents

Choose an example

Select a measured Cookbook case, then view its plan, progress, and revision or failure.

Simple PlanA multi-step plan is created, executed in order, and completed with no revisions.Plan ExecutionThe runtime tracks pending, in-progress, completed, and remaining steps against the plan.Plan RevisionAn observation makes the remaining plan inappropriate. Plan v1 is preserved; plan v2 continues.Plan FailureA required step cannot complete. The plan fails and terminates with plan_failed.

This Lab replays a measured Cookbook run; no API call is made.

Choose an example, then view the recorded Cookbook case.

Related Labs

The agent loop selects the next action; this Lab introduces an explicit plan the runtime can track and revise.

03 Agent Evaluation04 Planning

Continue Learning

Guides that explain planning, loops, and evaluation.

Implementation

Run the measured Cookbook example locally.

Agent Planning Cookbook

Related Tools

Tools commonly used with agent planning runtimes.

Related Rankings

See how this stack compares in DataAIHub rankings.

Best AI Agent Frameworks

Lab ID agent-planning · guided / precomputed