Start with Auto
Let Oppermind choose a quick response, deeper reasoning, or a Workspace Agent when you do not need manual control.
✦ Guide · Beginner
Choose the right chat mode and reasoning depth
Use Auto, Workspace Agents, Code agent, Browser operator, or Desktop co-pilot - and match reasoning effort to the consequence of the task.
Start guideYOUR OUTCOME
You will choose a mode and effort level based on the work - not on guesswork.
Use Auto as the default Choose an agent for actions and artifacts Increase reasoning for ambiguity or consequence
2 min
Choose a Workspace Agent
Use 1.0 for lighter everyday work and 2.0 for complex multi-step work that needs tools and in-app editors.
Use specialist control modes
Choose Code agent for multi-file apps, Browser operator for a live page, or Desktop co-pilot for installed native automation.
Set reasoning effort
Choose Auto, Min, Std, Deep, or Max according to ambiguity, complexity, and the cost of being wrong.
Review before consequential action
More reasoning can improve analysis, but it never removes the need for evidence, permissions, or accountable review.
What each lesson covers
The guide, lesson by lesson.
Start with Auto
Let Oppermind choose a quick response, deeper reasoning, or a Workspace Agent when you do not need manual control.
Leave the mode chip on Auto and type: Summarise the attached 40-page tender response into a one-page brief for the operations manager, then list the three clauses that need a legal check.
Choose a Workspace Agent
Use 1.0 for lighter everyday work and 2.0 for complex multi-step work that needs tools and in-app editors.
Click the mode chip that reads Auto and choose Workspace Agent 2.0, then ask: Build the onboarding pack for new warehouse casuals from the attached site rules: a welcome document, a first-week checklist sheet and a six-slide safety induction.
Use specialist control modes
Choose Code agent for multi-file apps, Browser operator for a live page, or Desktop co-pilot for installed native automation.
Click the mode chip and choose Code agent, then ask: Build a small booking web app for my dog-grooming business: a calendar page, a booking form with email confirmation, and an admin list of upcoming appointments.
Set reasoning effort
Choose Auto, Min, Std, Deep, or Max according to ambiguity, complexity, and the cost of being wrong.
Click the mode chip, set the EFFORT bar to Min for a quick job like reformatting a list of dates, then set it to Deep before asking: Using the attached quotes and our cash-flow sheet, should we lease or buy the second delivery van? Show the assumptions behind each option.
Review before consequential action
More reasoning can improve analysis, but it never removes the need for evidence, permissions, or accountable review.
After a Max-effort run recommends dropping the two lowest-margin products, ask: Show me the exact rows and figures from margin-analysis.xlsx that support each recommendation, and list anything you assumed that is not in the file.
Keep going
Make the learning stick
Start with one real piece of work.
Open Oppermind beside the lesson, apply each step, and leave with something you can use.