Web app live · Mac beta v0.9.5

When the decision matters, one AI answer is not enough.

Manwe turns hard questions and existing plans into evidence-backed decision records: AI advisors debate, evidence gets checked, dissent stays visible, and you leave with risks, forecasts, and next actions.

Decision record

The output is not a chat answer. It is a record you can inspect later.

A Manwe run keeps the recommendation, evidence, disagreement, risks, forecasts, and next actions together so a second opinion can survive beyond the conversation.

Manwe decision record displayed on a tablet
Decision record view: verdict, confidence, evidence, dissent, and what would change the answer.
Decision record 83% confidence

Should a B2B SaaS company automate 30% of support with AI agents?

No. The 30% target is built on an unverified denominator. Clean the ticket data first, then pilot narrow automation on deterministic workflows.

Advisory panel
  • CFO
  • CX Lead
  • Ops Strategist
  • Compliance Counsel
  • Contrarian
  • Auditor
Disagreement map
Cost savings casecontested
Customer trust riskhigh
Compliance pathnarrow
Surviving assumptions

Support volume may contain 25-40% duplicate, re-contact, or misrouted tickets.

What changes the verdict

Clean data, legal clearance, and a 60-day pilot with CSAT and escalation kill switches.

30 / 60 / 90 day plan
Verdict

A direct recommendation with confidence, not a vague summary.

Evidence

Sources, seed material, and claims stay attached to the record.

Dissent

The strongest objections remain visible instead of disappearing into agreement.

Risks

You see what could make the answer fail and what to watch next.

Forecasts

Pro runs add checkable predictions and Future Paths when the question supports them.

Next actions

The record ends with concrete moves, thresholds, scripts, or experiments.

What Manwe does

Bring the question or plan. Pressure-test it. Keep the record.

01 Bring a question or plan

Start with the choice, preferred option, stakes, constraints, and context the run needs to respect.

02 Advisors push back

Manwe casts role-bound advisors so the plan is challenged from more than one perspective.

03 Evidence gets checked

The run searches, ranks sources, tests fragile claims, and keeps counterarguments alive.

04 You leave with a record

The final artifact shows the answer, why it holds, what could change it, and what to do next.

Choose your surface

Use the web app for fast runs. Use the Mac beta when you want more control.

Both surfaces create decision records. The web app is the clean starting point. The Mac beta adds local-first depth, governed memory, scenario testing, and model flexibility.

Web app

Fast cloud runs

The quickest way to start: Quick and Pro runs, saved records, seed material, same-language recordings, and Pro Future Paths.

Start in web app
Mac beta

v0.9.5 local-first depth

Go deeper with Worlds, governed memory, What Manwe sees, scenario playgrounds, Lifeline, local models, cloud models, and custom APIs.

Download Mac beta
Web app Start a decision
Manwe web app start screen for creating a decision record
Use the web app when you want the fastest path from a hard question to a saved record.
Mac beta World memory view
Manwe Mac beta Worlds and Lifeline view
Worlds, memory review, scenarios, and Lifeline make context inspectable before it shapes a run.

Futures and predictions

Good decisions need a view of what could happen next.

Manwe does not stop at a recommendation. Pro runs can add forecasts, Future Paths, and a Deeper Story so the decision is tested against time, assumptions, and second-order effects.

  1. 01 Forecasts

    Serious records can include checkable expectations so the decision has a reality test.

  2. 02 Future Paths

    Manwe projects divergent timelines from the evidence, not just from the fear in the question.

  3. 03 Deeper Story

    Causal Layered Analysis surfaces the assumptions and narratives underneath the practical answer.

Endorsed by the creator of Causal Layered Analysis

Brilliant stuff here. Just love it. You are on to something. Congrats.

Prof. Sohail Inayatullah UNESCO Chair in Future Studies at IIUM Malaysia, creator of Causal Layered Analysis

Manwe implements Inayatullah’s Causal Layered Analysis and the Deeper Story synthesis. That matters because the product does not stop at advice: it studies the assumptions underneath the decision and the futures the decision could open.

Memory you control

Manwe remembers context only when you let it into the room.

The Mac beta v0.9.5 adds Worlds, governed memory, What Manwe sees, scenario playgrounds, and Lifeline. Suggestions stay out. Drafts stay out. Only promoted Memory can shape future runs.

01

Worlds

Keep work, personal, market, client, or project contexts separate instead of blending every decision together.

02

Memory review

Suggestions stay out. Drafts stay out. Only promoted Memory can shape future runs.

03

What Manwe sees

Inspect the profile, selected worlds, promoted memory, and relationships before a run starts.

04

Scenario playground

Test assumptions, ignore memory, or force outcomes without changing your real world context.

Manwe Mac beta Scenario playground with branching decision assumptions
Scenario playground: temporary assumptions and forced outcomes stay separate from promoted memory.

Public proof

Inspect real decision records before you run your own.

Curated examples show how Manwe handles tradeoffs, evidence, disagreement, risks, forecasts, and action plans across serious and everyday decisions.

Where it helps

Use Manwe before decisions that cost money, time, trust, or regret.

01

Should I take this job?

Compare money, trajectory, risk, identity, and what you may regret later.

02

Should we automate this workflow?

Test cost, trust, human-in-the-loop boundaries, and what success should actually mean.

03

Should I invest, wait, or exit?

Use evidence, scenarios, and explicit reasons that would change the answer.

04

I have a plan. What could make it fail?

Use Manwe as a serious second opinion before the plan becomes an expensive surprise.

05

Which family decision is least bad?

Slow down emotional tradeoffs and leave with a plan you can explain.

06

What happens next if we choose this path?

Use Future Paths to see what the decision could open over time.

FAQ

Plain answers.

How do I try Manwe?

Start in the web app for the fastest path from a hard question to a saved decision record. Download the Mac beta if you want local-first workflows, Worlds, governed memory, scenario testing, and more model control.

Is Manwe a report service?

No. Public records are proof of the output. The product is the process that creates private decision records you can act on, export, and revisit.

Who is it for?

Manwe is for people making decisions with consequences: founders, operators, consultants, product leaders, investors, professionals, and families facing hard tradeoffs.

When should I use Manwe instead of normal chat?

Use normal chat for quick explanations, drafts, summaries, and lightweight work. Use Manwe when the answer or plan needs pressure, evidence checks, dissent, and a record you can defend later.

Can Manwe give a second opinion on an idea I already have?

Yes. Manwe is useful when you already have a plan, option, or instinct and want it challenged before you commit time, money, trust, or reputation.

Does it replace professional advice?

No. Manwe is decision support. It helps inspect assumptions, evidence, options, and risks; it does not replace qualified legal, medical, financial, or professional judgment.

What is different about the Mac beta?

The Mac beta adds deeper local-first workflows: Worlds, governed memory, What Manwe sees, Lifeline, scenario playgrounds, local models, cloud models, and custom APIs.

What file formats can I export?

Decision records can be exported as PDF, HTML, or Markdown, including the verdict, evidence, dissent, risks, forecasts, action plan, and source split.

Start

Bring Manwe one decision you keep circling.

Use the web app when you want a fast decision record. Use the Mac beta when the decision needs local-first context, memory you control, or deeper scenario work.