Using Zcrafter

Zcrafter is designed to let you move between normal AI chat and live mainframe work without changing tools.

Start with a clear goal

Good sessions usually begin with one of these:

• a question you want answered
• a file, dataset, or member you want reviewed
• a job you want submitted, checked, or cleaned up
• a group of related chats you want to organize into a project

Choose the right model

For most everyday work, start with Aura.

Use a heavier model when:

• you want a longer answer
• you are comparing options in detail
• you want a more substantial draft
• you need deeper reasoning than a fast answer

If you are unsure whether a model is currently available to you, open Settings > Usage.

Use chats for day-to-day work

You can use Zcrafter for:

• explanations and tutoring
• study guides and quizzes
• writing and rewriting
• planning and summarization
• mainframe-oriented questions and actions

Use projects to stay organized

Projects help you group related chats together.

Projects are useful when you are:

• preparing for an exam
• tracking a modernization effort
• debugging repeated job failures
• keeping several related dataset investigations in one place

If you are signed out, projects are limited. Sign in if you want to save and organize work long term.

Use Settings to control the experience

Important sections inside Settings:

• Account: See your sign-in state and plan.
• Usage: Track daily and monthly availability.
• Data controls: Review how you want to manage your workspace behavior.
• Personalization: Adjust the feel of the UI and response style.
• About: Find product links and support resources.

Connect your mainframe only when you need live actions

You do not need a live mainframe connection for normal chat. Connect your environment when you want Zcrafter to work on real:

• datasets
• members
• USS files
• JCL jobs
• spool output

Treat destructive requests carefully

If you ask Zcrafter to delete, cancel, or purge something, make the target clear. This is especially important for jobs, output, and cleanup actions.

Examples:

• Delete job JOB12345.
• Cancel my latest job.
• Purge all of my jobs from today.

Use follow-ups naturally

Once a result is on screen, follow-up prompts usually work best when they point at that context directly.

Examples:

• Explain that condition code.
• Open that member.
• Delete that job.
• Summarize the errors only.

When something feels off

Check these first:

• Settings > Usage if the issue is about model access or daily availability
• Troubleshooting if the app looks wrong or a workflow is stuck
• Interactive Mode if Zcrafter did not interpret your request the way you expected