Skip to main content
Rollout sessions let you rerun long-running agents from the exact point you care about, without leaving Laminar.

​
Why rollouts

When you are debugging long-running agents, three things slow you down:
  • You constantly context-switch between windows to edit code, run it, and inspect execution.
  • You wait a long time for the agent to reach the event you care about.
  • The agent can take a different path and never hit that event at all.
Rollout sessions solve this by letting you cache up to a span, tweak configuration (like the system prompt) in the UI, rerun from there, and inspect the new trace in the same page. Once the CLI is connected, you never have to leave the page.

​
Setup

  1. Wrap your main entrypoint with observe and set rolloutEntrypoint: true.
import { observe } from '@lmnr-ai/lmnr';

export const generateStream = observe(
  { name: 'generateStream', rolloutEntrypoint: true },
  async (messages, model, modelId, reasoningEffort) => {
    // Your agent code here.
  }
);
  1. Start the dev orchestrator: npx lmnr-cli dev route/to/entrypoint.js.
  2. Open the session from Rollout sessions in your project, or click the link printed by the CLI.
If the file exposes multiple rollout entrypoints, pass --function to select one. Set LMNR_PROJECT_API_KEY or use --project-api-key to connect.

​
Use in Laminar

  1. Open Rollout sessions and select your session.
  2. Pick a view (Tree, Timeline, or Reader). Use Search, Filters, and Metadata as needed.
  3. Click Cache until here on an LLM span to reuse earlier outputs and jump ahead on the next run.
  4. Edit Parameters or System Prompts in the left sidebar.
  5. Click Run. New spans stream in and you can inspect them immediately.
  6. Click Stop to cancel a run.
You can edit your code while the CLI is running. Hot refresh keeps the rollout session alive so you can continue debugging without leaving Laminar.