> ## Documentation Index > Fetch the complete documentation index at: https://docs.browserbase.com/llms.txt > Use this file to discover all available pages before exploring further. # Generate a script > Turn a completed Agent run into a reusable Stagehand script with the Optimize tool. Use the Optimize tool to turn a completed Agent run into a reusable [Stagehand](https://docs.stagehand.dev/) script. Optimize reads the run's steps and generates code that follows the same workflow. This works best after your Agent completes the task successfully. The completed run gives Optimize a working path to translate into code. ## Generate the script Open the [Browserbase dashboard](https://www.browserbase.com/overview), select **Agents**, and choose the completed run you want to reproduce. Completed Agent run with an arrow pointing to the Optimize button Click **Optimize** in the run view. Optimize uses the run's event history, so you don't need to describe every action again. Enter this prompt in **Ask about this session**: ```text theme={null} Can you create a Stagehand script to do the same thing as this Agent run? ``` You can also select **Write a script** from the starter prompts. Optimize panel with an arrow pointing to the Write a script starter prompt Optimize returns a Stagehand script based on the run. Review the URLs, actions, extraction logic, and expected result, then copy the code into your project. Treat the generated script as a starting point. Run it against the same workflow before you use it with new inputs or sites. Optimize response with an arrow pointing to the generated script Follow the [Stagehand quickstart](/welcome/quickstarts/stagehand) to install Stagehand, set `BROWSERBASE_API_KEY`, and run the generated file. Stagehand automatically uses [Model Gateway](/platform/model-gateway/overview) when you use a Browserbase-hosted browser without a provider-specific API key. You can run the generated script with one Browserbase API key and no separate model provider account. ## When to generate a script Generate a script when you have a successful Agent run and want more control over how the workflow executes. A script lets you: * Edit individual Stagehand or Playwright steps. * Add the workflow to an existing application. * Test changes with your development tools. * Reuse a known path for repeated tasks. Keep using an Agent when the workflow needs to adapt across sites, layouts, or tasks that you can't define ahead of time. ## Next steps Install Stagehand and run your generated script Use models through your Browserbase API key Improve an Agent from its run history Trigger Agent runs from your application