Ask a hard question
Start with the decision, what is at stake, the options you are considering, and any context Manwe should respect.
How it works
From hard question to decision record.
Manwe is built for decisions where a smooth chat answer is not enough. It frames the choice, casts advisors, checks evidence, keeps dissent visible, and writes the record you can revisit later.
Manwe casts advisors
The run gives different AI advisors different jobs so the answer is not coming from one undifferentiated perspective.
Evidence gets challenged
Manwe searches, ranks sources, checks fragile claims, and keeps dissent visible before the final answer is written.
You get a decision record
The result is a durable record with verdict, confidence, evidence, dissent, risks, forecasts, Future Paths, and next actions.
What a record contains
Enough structure to act, and enough uncertainty to stay honest.
Inside a run
One question becomes several jobs.
Manwe does not ask one model to answer as if every perspective were already covered. It separates the work into framing, research, advisor debate, audit pressure, synthesis, and action planning.
That is why the output reads differently from a normal chat reply. The recommendation is first, but the record also keeps the reasons it might be wrong, the evidence it depends on, and the next step that should happen outside the app.
Read curated examples or start with the plain-language explainer: what is a decision record?
A useful boundary
Use Manwe when the answer needs to survive inspection.
Normal chat is fine for quick explanations, drafts, summaries, and lightweight work. Use Manwe when the choice needs evidence, disagreement, risk checks, forecasts, or a record you may need to defend later. For deeper local-first context, compare the Mac beta; for common decision types, see Manwe use cases.
FAQ
How Manwe works, plainly.
What happens after I submit a decision?
Manwe frames the decision, casts role-bound advisors, gathers or ranks evidence, challenges weak claims, preserves dissent, and writes the result as a decision record with verdict, confidence, risks, and next actions.
How are advisors selected?
Advisors are chosen around the decision type and the jobs the run needs. A business automation question gets different lenses than a family, career, market, or technical decision.
Can I inspect the reasoning?
Yes. The record keeps the verdict, advisor positions, assumptions, evidence, dissent, risks, forecasts, and action plan together so the result is not just a final answer.
When should I use Quick or Pro?
Use Quick when you need a structured decision record fast. Use Pro when the question needs deeper claim checks, forecasts, Future Paths, or more explicit uncertainty.
Is the Mac beta required?
No. The web app is the main self-serve path. The Mac beta is for deeper local-first workflows, governed memory, Worlds, scenario testing, and more model control.