Get started
Chat modes and reasoning effort
Choose how Lato responds, acts, builds, browses, or controls installed desktop tools.
On this pageAutoWorkspace AgentsSpecialist control modesReasoning effortReasoning is not verification
Auto
Auto is the default for most work. It can choose a quick response, deeper reasoning, or a Workspace Agent according to the request.
Ask Oppermind anything…
AutoAuto
Try this
✦Leave the mode chip on Auto and ask:
What is the difference between an ABN and an ACN, in two sentences?
What you getOppermind answers directly in chat without planning steps or tools, because Auto judged this a quick question; the status on the right of the composer still reads Auto.
Workspace Agents
Workspace Agent 1.0 is the lighter, faster option for everyday tasks that need planning, tools, or artifacts. Workspace Agent 2.0 is for more complex work that needs a fuller agent and direct in-app editor access.
Ask Oppermind anything…
Workspace Agent 2.0Deep
Try this
✦Click the mode chip and choose Workspace Agent 2.0, then ask:
Turn the attached quarterly-results.xlsx into a six-slide board update: one slide per business unit, a summary slide and a risks slide, using only the numbers in the sheet.
What you getOppermind plans the deck, reads the spreadsheet, and the presentation opens straight in the Slide editor in the right-hand pane, where you can ask for changes slide by slide.
Specialist control modes
Code agent builds multi-file web apps and projects with files, terminal, preview, and shipping workflows. Browser operator controls or reads the live page when available. Desktop co-pilot handles installed native or desktop automation.
Basics
Auto~2–30sPicks the best approach for your request — instant for quick questions, deeper reasoning or a Workspace Agent when the work needs it.
Doing (agents & control)
Workspace Agent 1.0~30s–3mLighter, quicker agent — plans, uses tools, and produces artifacts for everyday tasks.
Workspace Agent 2.0~1–5mFor more complex work — full agent with desktop editor access; plans, runs tools, and delivers documents, sheets, decks and designs straight into their editors.
Code agent~1–10mBuilds web apps and multi-file projects — writes/edits files, runs a terminal, previews and ships code.
Browser operator~1–10mOperate a specific website: open pages, click, type and fill forms. Not for normal chat or general search.
Desktop co-pilot~1–10mNative/desktop automation path (when installed).
Efforthow hard Lato thinks
AutoMinStdDeepMax
Ask Oppermind anything…
AutoAuto
▢New conversation ▾⌕
Basics
Auto~2–30sPicks the best approach for your request — instant for quick questions, deeper reasoning or a Workspace Agent when the work needs it.✓
Doing (agents & control)
Workspace Agent 1.0~30s–3mLighter, quicker agent — plans, uses tools, and produces artifacts for everyday tasks.
Workspace Agent 2.0~1–5mFor more complex work — full agent with in-app editor access; plans, runs tools, delivers documents, sheets, decks and designs into their editors.
Code agent~1–10mBuilds web apps and multi-file projects — writes/edits files, runs a terminal, previews and ships code.
Browser operator~1–10mControl and read the live web page (when available).
Desktop co-pilot~1–10mNative/desktop automation path (when installed).
Efforthow hard Lato thinks
AutoMinStdDeepMax
Try this
✦Switch the mode chip to Code agent and ask:
Build a small web app for our gym's class bookings: a list of classes with capacity, a booking form, and an admin page to add classes. Start with a simple front end and mock data.
What you getOppermind creates a multi-file project in the Code editor with a file tree, terminal and live preview, and you steer each feature before deciding whether to deploy.
Reasoning effort
Choose Auto, Min, Std, Deep, or Max. Increase effort when the problem is ambiguous, multi-step, difficult to reverse, or costly to get wrong; use lower effort for simple transformations and quick questions.
Efforthow hard Lato thinks
AutoMinStdDeepMax
Efforthow hard Lato thinks
AutoMinStdDeepMax
Try this
✦Click the mode chip that reads Auto, set the EFFORT bar to Min, then ask:
Convert this pasted list of 40 supplier names to Title Case and remove the duplicates.
What you getThe cleaned list comes back almost immediately; Min suits a mechanical transformation where extra reasoning would only add waiting time, and the status on the right shows the effort you picked.
Reasoning is not verification
A deeper setting does not prove that a claim is correct or an action is safe. Check evidence, permissions, calculations, and consequential decisions independently.
You said our break-even is 1,840 units a month. Show the exact formula and every input you used, and point to the cell in the attached costing.xlsx that each input came from.
◆ Lato · Pro
Break-even = fixed costs / (price - variable cost per unit).
Fixed costs $46,000 a month (Costs!B12); price $62 (Pricing!C4); variable cost $37 per unit (Costs!B20).
46,000 / 25 = 1,840. Check those three cells in the sheet before it goes in the deck.
Fixed costs $46,000 a month (Costs!B12); price $62 (Pricing!C4); variable cost $37 per unit (Costs!B20).
46,000 / 25 = 1,840. Check those three cells in the sheet before it goes in the deck.
Try this
✦You said our break-even is 1,840 units a month. Show the exact formula and every input you used, and point to the cell in the attached costing.xlsx that each input came from.
What you getOppermind lists the formula, inputs and source cells in chat, and you open the Sheet to confirm each cell matches before the number goes into the investor deck.
◎
GUIDED LEARNING
Open learning path Practise this in Oppermind Academy
Follow the related tutorial or course and apply the concept to a real task.