Give your AI brain mental models for better decision-making. Not content you read, but infrastructure you run. Each model is an executable skill that walks you through applying it to real decisions.
One mental model gives you a lens. Chaining models gives you a decision. Each model's output feeds as context into the next, building compound insight that no single model could produce alone.
No menus. No dashboards. Type a command, describe your decision. The brain does the rest.
Describe your problem in plain language. The brain analyzes it, recommends 3-5 relevant models, runs them in sequence and produces a synthesized decision brief with confidence level.
Manually compose a chain of models to run against a decision. Each model's output feeds as context into the next. You control the exact sequence.
Run multiple models on the same problem side-by-side. See where they agree, where they conflict and what each one surfaces that the others miss.
Browse the full model library with search and filters. Find models by category, difficulty, decision type or chaining compatibility.
Each chain runs 4-6 models in sequence against your specific problem and produces a decision brief. Run them as-is or customize the model sequence.
| Chain | Model Sequence | Use When |
|---|---|---|
| Pricing Decision | anchoring → scarcity → supply-demand → trade-offs → second-order | Setting, raising or restructuring pricing |
| Market Entry | first-principles → customer-dev → network-effects → bottlenecks → defensibility → trade-offs | Entering a new segment or geography |
| Product Launch | first-principles → mvp → assumption-mapping → experiment-design → pmf → second-order | What to build, what to skip, how to validate |
| Hiring Decision | circle-of-competence → inversion → incentives → second-order → trade-offs | Who to hire and what would make the hire fail |
| Investment Decision | probabilistic → margin-of-safety → inversion → second-order → sunk-cost → trade-offs | Evaluating time, money or resource investments |
| Should I Quit | sunk-cost → inversion → local-maxima → pivot-or-persevere → first-principles | When to quit, pivot or double down |
| Life Decision | first-principles → inversion → second-order → trade-offs → regret-minimization | Career moves, relocations, life changes |
Claude Code is brilliant at following instructions. But following instructions is not the same as making good decisions. The gap between execution and judgment is where most AI-assisted work falls short.
It follows your instructions to the letter but doesn't question whether the instructions are right. You get exactly what you asked for, which is a problem when you asked for the wrong thing.
It never says "have you considered the opposite?" or "what if this assumption is wrong?" It accepts your framing and optimizes within it, even when the framing is the mistake.
It optimizes what you ask for but doesn't ask if you're optimizing the right thing. First-order effects are covered. Downstream consequences, blind spots and trade-offs are invisible.
Mental Models OS is not a static library you read. It is a loop. Every decision you run through it leaves a trail. The next decision starts with the prior trail as context. By month three, the brain knows your decision biases, your common chains, your usual confidence levels. Decision quality compounds the way memory compounds in any second brain.
/think on a hard call/learn logs pattern, next call is smarterSession 1. You type a question. Brain picks 5 models cold. Decision brief comes back at 65% confidence. You log it. That is one data point.
Session 10. Brain has seen your past 9 decisions. It recognizes your decision patterns and your blind spots. Confidence climbs to 75%. Chains run faster because the brain knows your usual frame.
Session 50. Brain has extracted reusable chains from your history. It surfaces past similar decisions as context. Confidence routinely above 85%. You stop second-guessing.
Every model is an executable skill. Not theory you read. Workflows your Claude Code runs against your real problem. Each one walks step-by-step application to the decision you typed.
New models added regularly from consulting engagements and community requests. Type /index to browse the latest.
The calls that come up most. Each chain runs 4 to 6 models in sequence against your specific input and returns a decision brief.
Most frameworks assume you have data. Early-stage decisions don't. Mental Models OS is built for the messy middle: when you're making bets with incomplete information and high stakes.
Test desirability first: does anyone want this? Then viability: can it make money? Only then feasibility: can we build it? Most teams go in reverse order. They build first and hope demand follows.
Rank your riskiest assumptions by impact and certainty. Kill the ones that would sink the business before you invest in building. The brain walks you through identifying hidden assumptions you didn't know you were making.
For every assumption, the brain recommends the cheapest possible test. An email before a landing page. A landing page before a product. A product before a platform. Validate demand at each step before investing more.
75 mental models. 7 decision chains. 4 meta-skills. Plus 100+ playbooks, 100+ skills, 100+ specialized agents. 239 operators have already installed it. Install once, reason better forever.